The Importance of Forgetting

Abstract Morality bears on what we should forget. Some aspects of our identity are meant to be forgotten and there is a distinctive harm that accompanies the permanence of some content about us, content that prompts a duty to forget. To make the case that forgetting is an integral part of our moral duties to others, the paper proceeds as follows. In §1, I make the case that forgetting is morally evaluable and I survey three kinds of forgetting: no-trace forgetting, archival forgetting, and siloing. In §2, I turn to how we practice these forms of forgetting in our everyday lives and the goods these practices facilitate by drawing on examples ranging from the expunging of juvenile arrest records to the right to privacy. In §3, I turn to how my account can help us both recognize and address a heretofore neglected source of harm caused by technology and big data. In §4, I end by addressing the concern that we lack control over forgetting and thus can't be required to forget. I argue this challenge can be answered, but there's a harder challenge that can't. Forgetting is under threat. To address this challenge and preserve forgetting, we must change the world.

I wish I could remember my grandmother's voice. She died half a lifetime ago, and I only have a few memories left of her. What I associate most with her is the sound of the static crackle of the television while she rubbed turmeric into my skin. She loved to watch Australian soap operas, and I would sit and watch with her. She couldn't speak a word of English, but she never once asked me to explain what was happening. With each day I feel more of my memories of her slipping away. I'm afraid I'm forgetting her. Then there are the memories I wish I could forget, like that time I was accidentally unmuted during a faculty meeting or similarly cringeworthy moments that keep me up at night. It feels like a further insult to my grandmother that these memories are the ones that stick around long past their welcome, memories that feel just as vivid as the day they occurred. Then there are memories that I have been asked to forget. I would tell you about these if I could, but these memories have been wiped from the record.
Generally, forgetting is regarded as a bad thing. Realizing we forgot something is often marked with a feeling of loss, and whatever was forgotten is mourned. Despite the general intuition that forgetting is a negative detriment to our lives, forgetting can also play a positive role. I'm not the first to note this. For example, against George Santayana's warningthose who cannot remember the past are condemned to repeat it - Krondofer (2008) argues that forgetting can be a positive force. Krondofer points us to Jorge Luis Borges's short story about Ireneo Funes, a man who cannot forget. Funes's perfect memory comes at a cost, it traps him in the past. This short story points to an important lesson about forgetting: "without forgetting, the human species would have to relive the past continuously and never live in the present moment. Without forgetting, there would be no future" (Krondofer 2008: 239). Forgetting has been unfairly maligned in favor of remembering, but remembering can be just as detrimental to our lives. 1 Forgetting is an integral part of our moral duties to others. Some aspects of our identity are meant to be forgotten. That is, the practice of identity shaping depends on our ability to grow and change over time and a barrier to that growth and change is a failure to recognize that we may no longer be the same people we were before. There is a distinctive harm that accompanies the permanence of some content about uscontent that impinges upon our identity shaping practicesthat prompts a moral duty to forget. This distinctive wrong is the violation of the fundamental interest we each have in controlling, shaping, and maintaining our identity and how we present ourselves to others. 2 This fundamental interest, as I have argued elsewhere, extends to what others believe of us. 3 This fundamental interest, as I will argue here, also extends to what others forget about us. To make the argument, the paper proceeds as follows. 1 Michaelian (2011) also makes a positive case for forgetting in a similar vein. Michaelian argues that there are patterns of forgetting that can aid in improving the overall shape of the agent's total doxastic state. That is, our human limitations with regard to memory should not be considered unfortunate, but rather a helpful aide. Forgetting too much can be a vice, so too can remembering too much be a vice. Following an Aristotelian approach then, what we need to find is the mean between those opposed vices, that is, the right amount of forgetting. Along a different vein, Singer et al. (2021) offer an argument for the importance of forgetting but focus on how group learning outcomes can be improved not only by how agents form beliefs, but also how they forget them. In addition, with regard to remembering too much, see the discussion of hyperthymesia as it relates to our ability to forgive in §2.2.
In §1, A Framework for Forgetting, I present my account of doxastic wronging and demonstrate how it can function as a framework for identifying when forgetting is morally evaluable. I will not be defending doxastic wronging, but rather demonstrating how the theory applies to forgetting to illuminate what is morally valuable about forgetting. With this framework laid out, I can also distinguish three forms that forgetting can take: no-trace forgetting, archival forgetting, and siloing.
In §2, The Goods Forgetting Facilitates, I turn to how we practice these forms of forgetting in our everyday lives and the goods these practices facilitate. For example, the practice of expunging juvenile arrest records is motivated by a desire for no-trace forgetting. The expunging of records enables an erasure that allows one's identity to change and grow so that one's future is not improperly constrained by the past. Similarly, our practices of forgiveness involve forgetting. Forgetting helps to ease the process of forgiving to facilitate growth and change over time. We can think also of privacy violations and how such violations engender duties to forget. And finally, many of our norms of etiquette, norms that help ease social interaction, involve not just practices of remembering but also practices of forgetting. Central to all of these practices of forgetting is the role that forgetting plays in another practice of significant moral value, our practice of shaping identity.
With the goods that forgetting facilitates laid out, in §3, A New Threat to Forgetting, I turn to how the account developed thus far can help us both recognize and address a heretofore neglected source of harm caused by technology and big data. The internet never forgets. The ease of access to private information, the permanence of that information, and the incentive structures that push companies towards collecting more and more information threaten the goods outlined previously. The European Union's General Data Protection Regulation (GDPR) introduces a right to be forgotten to combat this new threat but there is a deeper threat to our identities that my account of forgetting can bring to the surface.
Finally, in §4, Two Control Worries for Forgetting, I turn to a familiar worry for any account that suggests we have obligations with respect to our attitudes. This worry follows from the intuitive thought that ought implies can. That is, if you ought do a thing, you must be able to do that thing. Forgetting, much like belief, is commonly thought to be out of our control. If that's so, then the argument of the paper amounts to naught. I argue that we can be obligated to forget because forgetting is a process that we exercise (some) control over. This worry, however, points to a harder worry about control presented not by our own volitional abilities, but rather by the world itself: often the world can act as an uncontrollable impediment to forgetting. Although this is a harder challenge, it reinforces why forgetting is important. A world without forgetting is a world in which we're unable to secure our fundamental interest in controlling, shaping, and maintaining our identity and how we present ourselves to others. Forgetting is a right that must be preserved. If the world stands in the way of that right, then the world must be changed.

A Framework for Forgetting
We often seek apology not just for what people say or do to us, but also what they believe of us. 4 For example, suppose your mother believes you haven't made good 4 I will not rehearse all the arguments for doxastic wronging here, focusing instead on the general motivations and framework. For more arguments see Basu and Schroeder (2019), Basu (2019aBasu ( , 2019bBasu ( , 2021. For extending doxastic wronging beyond belief see Basu (Forthcoming-a) and Basu (Forthcoming-b). use of your degree. 5 We can note three things about this belief. First, the belief is directed. The belief is about you. Second, although it might be more hurtful if your mother reveals this belief about you, e.g., while complaining about you at a dinner party, the wrong she commits against you is in the belief itself. To put it simply, she shouldn't have the belief. And finally, she wrongs you in virtue of the content of what she believes. You have a legitimate complaint about what she believes of you. 6 In Basu (2021), I argue that many of these doxastic wrongs can be understood as holding failures. The notion of a holding failure comes from Lindemann (2016) where Lindemann argues that personhood is a socially shared practice that involves two aspects: holding and letting go. According to Lindemann our identities are narratively constituted. That is, the answer to the question "Who am I?" is constituted by the stories we, and others, tell about ourselves. Holding involves a kind of storytelling, and it begins at home. As I've noted previously: The centerpiece of Lindemann's theory of holding is that our identities are pieces of narrative construction that are constituted in part by others' beliefs about us (as well as our beliefs about ourselves), and thus, we fundamentally depend on others. Holding, when done well, "supports an individual in the creation and maintenance of a personal identity that allows her to flourish personally and in her interactions with others." 7 Holding, when done badly, can be destructive. (Basu 2021: 110) We have a moral obligation to hold others well, and this extends beyond our actions regarding them, it extends to how we think of them. Given the importance of how others hold us in their thoughts, there's an obligation to perform this holding well.
The identity work at the center of our interpersonal relationships involves not only believing certain things about our loved ones, but also forgetting. In contrast to holding there is letting go. According to Lindemann (2016: 85), letting go is the process of "weeding out the stories that no longer fit and constructing new ones that do." If my parents continued to think of me in line with the stories they had constructed of who they thought I'd be before I came out that'd be a way they can wrong me because their stories no longer accurately reflect who I am. 8 In an important sense, they need to forget it. It's in the past. We ought to let go of our stories about people when those stories no longer fit who they are. It is here, at letting go, that we get a duty to forget. Forgetting plays an important role in this kind of mutual recognition that's central to the relationships we stand in with one another. 9 Notice now that letting go, or forgetting, can also demonstrate the three hallmarks of doxastic wronging that I outlined earlier. First, the kinds of forgetting that are morally evaluable are directed. The kinds of morally evaluable forgetting that are of interest in this paper are those that are about other people, ways of holding, or not holding, others in our thoughts. Second, forgetting is morally evaluable as forgetting. That is, although we might seek apology for other things that follow from this act of letting go, e.g., 5 This example is from Basu (2021: 107-8). 6 That is, assuming you disagree with your mother's assessment of you. 7 Lindemann (2016: x). 8 For more on how parental expectations can wrong see Basu (Forthcoming-a). 9 In addition, there is also an epistemic dimension to this moral forgetting. When it comes to understanding who a person is now, old evidence will have less predictive power, it'll be less accurate. So, if you hold onto old evidence about a person, you're more likely to be led astray by those outdated stories.
continually bringing up narratives that no longer fit, the initial complaint in cases where one should've forgotten but didn't is about the forgetting itself. Third, there is some content to the state in question, there is some content that needs to be forgotten, to let go. Furthermore, forgetting isn't just something that happens to us, forgetting is a process that we frequently guide. 10 To better understand forgetting and how it can be morally evaluable let's consider some examples.
With respect to the small stuff, we tend to think that our memory functions to track importance. It's not only harder to forget a friend's birthday, but it's also worse when we do, because of how important friends are to us. 11 When a student forgets a meeting that they had scheduled with me, it hurts because it means that I'm unimportant to them. However, it doesn't hurt nearly as much as that time my parents forgot to pick me up from school. As Robert, in the Netflix show Grace and Frankie, remarks to his husband when he's questioned about his recent bouts of forgetfulness, "I'm going to let you in to a little secret about my brain. I have an important file and an unimportant file and the latter is where things go that aren't important to me in the moment." 12 Unfortunately, as is revealed later, Robert's memory is in fact deteriorating, and his husband was right to be concerned. Generally, however, a low-stakes meeting for coffee or the answer to the question of what you ate the previous day are not as important to us, and thus, they are easier to forget.
Further, it often makes sense to say that you forgot a thing where you still have the ability to recall it. For example, when you know that you know that actor from something you've watched recently but just can't place them, but while washing the dishes later that evening their name comes to you. When we say we've forgotten in this way what we've really done is archived. Think back to Robert's process of sorting information into one of two files, either the thing is important or it is unimportant. 13 Robert's classification system functions as a form of forgetting. When you place something in the unimportant file the item placed there can still be recalled, but not as easily or quickly as items that are placed in the important file. As items fade away into the background or are archived away, they function as forgotten. Let us call this form of forgetting archival forgetting. 14 That is, with time what's in the unimportant file fades from our minds. 15 These sorting choices reflect our values and moral judgements about what is important 10 Sometimes the forgetting is out of our control entirely, and terribly important things can be forgotten, e.g., as forgetting due to conditions like Alzheimers demonstrate. I will return to the challenge the question of control poses to whether we can be required to forget in §4.

11
See Smith (2005). 12 This conversation occurs in Season 7, episode 9, "The Prediction." 13 There is reason to be suspicious of the storage-space, computer, and filing metaphors that are being employed here. Memories are not perfectly preserved, to be recalled later when needed. When memories are recalled, there is a process of reconstruction, our memories get modified. Our memories can also be affected by how other people talk about our memories, and false memories can be implanted by photographs, videos, and even just people telling us that a thing that never happened to us in fact did. Thanks to Alex Madva for pressing me on these further complications, but I set them aside for the purposes of this paper. I hope to be able to return to them in future work because that memory operates in this way is more evidence of the importance of narrative and social shared identity and the work of holding and letting go that I discussed earlier.
14 I owe the language of this distinction between archival forgetting and no-trace forgetting to Jason Kawall's comments. 15 For example, do you know what emails are on your 137th page of emails? Likely not. Unless the email has been starred, flagged, or moved into an important folder of some kind, although it may technically be there, it has been forgotten. The other emails in your inbox exist, they're just archived in ways that diminish and what is not, and as such these choices can be subject to legitimate complaints from others. When we forget when we shouldn't have, because of these choices we've made, we can be held morally accountable. Now contrast archival forgetting with the cases of forgetting demonstrated by the epigraphs I've chosen for this paper. Both Lila and Clementine desire forgetting, but only Clementine comes close to achieving it. The first epigraph is Elena Greco discovering that her childhood friend, Lila, has vanished. Lila's possessions are gone, she has cut herself out of every picture, there is no evidence left behind that would suggest Lila used to live in her apartment. However, as much as Lila might try to disappear, as much of her past life she might try to destroy, there is evidence of her existence she can't erase. Namely, Elena remembers her. Driven by anger at Lila's disappearance, Elena denies Lila her wish to disappear. Elena begins writing about her friend. Elena will preserve Lila's story. The desire to have people forget us may not be familiar to everyone but this case brings out the messiness of holding and letting go. Lila desires to be forgotten, for all of her to be let go. Elena, however, refuses to forget her best friend. Unlike forgetting stories from the past that no longer fit the people we are, Lila goes too far in asking Elena to forget her entire existence. That's not a demand that Lila can make.
However, we often remove people from our lives in an effort to move forward, we try to forget them because otherwise our memories can keep us anchored in the past. In Charlie Kaufman's 2004 film, Eternal Sunshine of the Spotless Mind, Clementine wants to erase all the memories of an ex, Joel. The reason Clementine succeeds but Lila doesn't is twofold. First, the technology exists in Clementine's world to remove the memories we have of others. Second, Clementine only wishes to remove her memories of Joel, not for Joel to forget her. 16 However, even Clementine's case is complicated because the memories that she has erased are still recorded on tape. Thus, both cases also share an important similarity: in both cases we see a desire for no-trace forgetting, i.e., a total erasure of the memories of the person, but in both cases, we also see that desire for no-trace forgetting thwarted. In Lila's case it is because Elena still retains her memories of her. In Clementine's case it is because the company that erased her memories preserved a recording of her memories of Joel, and that copy is mailed to her in the closing act of the film.
The no-trace forgetting desired by both Lila and Clementine is likely impossible. Perhaps the closest we can get to no-trace forgetting in our actual world is the practice of damnatio memoriae as a form of punishment where all record of a person's existence is erased from history. Notice, however, that for this form of punishment to be enacted at least one person must still remember. You cannot direct the destruction of information without knowing what you're looking for. Even this kind of forgetting is a process. Furthermore, it is an active process unlike the passive process that occurs with items placed in one's unimportant file. To complicate matters further, the absences themselves record a kind of archive. Historians study those absences to figure out who might've been erased. 17 But, it's also worth noting that if there has been a successful version of no-trace forgetting, we wouldn't know. their importance. However, when you pull up the email it will be perfectly recalled. Our memory, unfortunately, doesn't work like that. When we recall we must also reconstruct as I discuss in fn 13. 16 However, upon learning that Clementine has undergone this procedure, Joel also elects to undergo it. 17 In comments Lauritz Munch (Ms, 2021) has suggested there might also be an interesting analogy here between deep and shallow secrets. Following Pozen (2010), deep secrets are government secrets known to a small group of officials and concealed from anyone outside of that small group, whereas shallow Drawing on the literature on classification systems and organizational structures, we can also shed light on the nature of forgetting and when we can be morally evaluated for how we structure our forgetting by looking at how different organizational structures can facilitate different kinds of forgetting. 18 For one example, Bowker and Star (2000) note the following positive cases of forgetting: when rediscovery is easier than remembering and when forgetting facilitates a change in identity. On the first form, Bowker and Star (2000: 257) use the example of how airline companies "do not retain a record of one's food or seating options" because the cost of constructing archives for such precise and large amounts of data are too high for the payoff, thus it's easier (and cheaper) to "process passengers anew each time they are encountered." For the second form, Bowker and Star (2000: 257) note that "when an organization wants to change its identity … the argument is that 'we have always done things this way' stands in the way of breaking new ground." This is the form of forgetting that's central to our practices of holding and letting go.
Bowker and Star also draw a distinction between clearance forgetting and erasure forgetting. They note that clearance forgetting is "the erection of a barrier in the past at a certain point so that no information or knowledge can leak through to the present" whereas erasure forgetting is "the ongoing destruction of selective traces in the present" (2000: 257). Erasure forgetting clearly resembles the example of damnatio memoraie and our discussion of no-trace forgetting. Clearance forgetting, however, suggests a third category of forgetting, one that is distinct from both no-trace and archival forgetting. Clearance forgetting would be more like if Robert had a third folder in addition to his important and unimportant folders, a folder for undesirables. This is a folder for memories he wanted to quarantine from other memories, memories he didn't want to play any role in his inferences. The items chosen for this folder are chosen on the basis of their content. However, the metaphor of a folder is too weak for the degree to which we wish to isolate these kinds of memories. These are the memories around which he's built walls, perhaps because the memory is just too painful or perhaps because they are memories that he's been asked by others to silo off from the rest of his mind. Running with this imagery, let's call this form of forgetting siloing. Thus, we can identify at least three forms of morally evaluable forgetting: erasure or no-trace forgetting, archiving or archival forgetting, and siloing or siloed forgetting. Further, in each of these cases, we can see the hallmarks of doxastic wronging. These forms of forgetting are directed, you're morally evaluable for the forgetting itself, and there is some content in question that you are being asked to forget or that you desire to forget. Now that we have some clarity on how forgetting can be morally evaluable and the forms that forgetting can take, I turn now to the question of when forgetting is desirable. That is, the positive role that forgetting can play in our lives and why, when, and how it ought to be encouraged. With regard to the case of organizational structures that facilitate both remembrance and forgetting, we have a large literature on organizational norms governing classification, access, and the storage of information because it's been recognized that these are questions which have moral and political consequences. As secrets are things hidden from the public but the public knows that they are being denied some items of information. Put more simply, with shallow secrets you know something's being withheld, but you don't know what. With deep secrets, you don't even know that something is being withheld. Pozen (2010: 257) suggests that "every act of state secrecy can be located on a continuum ranging between these two poles." Whether forgetting can be modeled analogously is a question I set aside for others. 18 Thanks to C. Thi Nguyen for suggesting I look at Bowker and Star's (2000) book for these points.
such, these questions must be answered diligently with these consequences in mind. A similar concern with respect to our individual practices of classification, access, and the storage of information has not been as widely recognized. By considering a number of case studies in the next section I will demonstrate how lessons from the goods that organizational forgetting both protects and facilitates also appear in the individual level.

The Goods Forgetting Facilitates
I turn now to some case studies with the aim of demonstrating the goods that forgetting helps facilitate and protect. As will be revealed, we already have practices of forgetting that suggest we're responsive to moral considerations bearing on the question of what we should remember and what we should forget. My goal is just to make the desirable features of these practices explicit so then we can subject them to closer scrutiny.

Juvenile Arrest Records
As Radice (2018: 365) writes, "many people believe that juvenile records are not public, especially after a juvenile turns eighteen." However, as she continues, "that common notion is a myth." Although this common notion has become a myth, it was not always so. The common misconception that juvenile arrest records are not public stems from the origins of the juvenile justice system. As Radice (2018: 369) notes, "juvenile courts were the first courts to expunge or destroy records, relying on the premise that juveniles should be able to outgrow their youthful indiscretion and be given a clean slate in adulthood." A key factor in the origins of juvenile courts was this idea of a separate juvenile court system in which the records would be erased in order to save "the child from the brand of criminality, the brand that sticks to it for life" (Radice 2018: 385). As the Supreme Court of Arizona wrote in 1965, "the policy of the juvenile law is to hide youthful errors from the full gaze of the public and bury them in the graveyard of the forgotten past." 19 With this background, we see in the origins of the juvenile court system a valuable good that no-trace forgetting facilitates. The forgetting facilitates a change in identity, the forgetting ensures that one's future isn't improperly constrained by one's past. This valuable good, this fresh start enabled by no-trace forgetting, has been undermined in recent years by racism. Faced with a public panic over rising crime-rates, juvenile record protections were eroded over the course of the 80s and 90s in the interest of protecting the public from a new class of criminals, so-called "superpredators" (see Funk 1996). Although debunked as a racist myth, this demonization of Black youth has undermined the goods that the juvenile justice system sought to facilitate through the expunging of records when juveniles became adults. 20 It is now the case that juvenile arrest records contain more information than ever beforephotographs, fingerprints, and often, also DNAand that information is shared even more widely. For example, the information is often shared with the juvenile's school, which could result in suspension or expulsion, which would then become part of their permanent record when applying to colleges. The information is often also shared with social services, and if the juvenile lives in public housing, they could be 19 In re Gault, 387 U.S. 1, 32 (1967) (quoting In re Gault, 407 P.2d 760, 767 (Ariz. 1965)). 20 See https://www.themarshallproject.org/2020/11/20/superpredator-the-media-myth-that-demonizeda-generation-of-black-youth for more on the "superpredator" myth. evicted as a result. Even if the juvenile is lucky enough to live in a state that still practices the expungement of juvenile arrest records, the record can exist in some form across a number of different agencies, each with their own policies about how that record will be handled. What they do with the records could lead to further action that is harder to erase from one's history. This dispersal of the records across different agencies also increases the risk that private information could become public, for a record to follow one even when it shouldn't. That is, for the silo to leak. Recall that not even Clementine could achieve no-trace forgetting because a recording of her memory of Joel was preserved. How can anything be buried in the graveyard of the forgotten past if records of the past still exist?

Forgiveness
When someone asks you to forgive them, they too seek an expungement of the past. Forgetting helps facilitate the process of forgiveness. Central to forgiveness is the idiom that to forgive is to forget. Like most sayings, the idiom obscures a number of nuances. For example, it is just as common to hear that one should forgive but not forget, as in the case of truly atrocious crimes where the forgetting of what happened would be a further insult. However, forgiveness does seem to require, on the part of the forgiver, some act of wiping the slate clean, of not holding onto past misdeeds, of expunging the past from their records. If someone has forgiven us for some wrongdoing we have committed in the past, we tend to think that it would be unjust if they continued to bring up the wrongdoing. 21 This suggests that once forgiven the misdeed must be archived or siloed away.
To fully forgive someone seems to involve some form of forgetting about the past. That is, you cannot truly forgive if the wrongdoing is still at the forefront of your mind. Forgiving is a process that helps to facilitate forgetting. 22 As a patient with hyperthymesia reports, forgiving is harder for him because he can never forget what 21 For example, when Monique Wonderly (2021: 7) canvasses much of the forgiveness literature she finds precisely this kind of desire for archiving or siloing the memory of the misdeed(s).
Jean Hampton claims that in forgiveness, "the victim 'sends away' the immoral action in the way that a creditor would absolve a debt" (Murphy & Hampton, 1988, 38). Christopher Bennett writes, "In forgiving, a person puts the wrong behind them in some way, determining to leave it out of consideration in their attitude towards the wrongdoer" (2003,127). On Lucy Allais's account, when you forgive an offender, you "wipe the slate clean" in the sense that "the [wrongdoer's] act is disregarded in your ways of regarding and esteeming her" (2008,56). 22 Both the issues of juvenile arrest records and what forgiveness requires are more complicated than the quick gloss that I have provided here. For example, although I draw on the survey of the literature from Wonderly (2021) in fn 21, she herself disagrees with the idea that forgiveness requires wiping the slate clean. In that paper, Wonderly (2021) argues that retracting forgiveness, i.e., un-forgiving, can be justified. Unforgiveness requires remembering. Furthermore, perhaps in the case of juvenile arrest records, there are some crimes, even if committed by juveniles, that should never be forgotten. So, although the erosion of protections surrounding juvenile arrest records was motivated by racism, there are some actions that should be preserved in the records forever. For a further complication, in contrast to a duty to forget there is also a duty to know (see Lackey Ms). The right to be known plays an important function not only with respect to identity and mutual recognition, just as the right to be forgotten does, but also the right to be known functions to correct past injustices, injustices that importantly, should not be forgotten. Although this suggests a tension between forgetting and knowing, as I'll argue later the duty to forget is a prima facie duty that can be defeated by other considerations. Sometimes the duty to know will take precedence. These are issues worth exploring further, but I do not have the space to take up in detail here. was done. 23 Forgetting facilitates a valuable good we find in both the origins of the juvenile court system and in our practices of forgiveness. Sometimes we must let go of the past.

The Right to Privacy
We have strong intuitions about cases in which someone's right to privacy has been violated. For example, take the paradigmatic case of a burglar breaking into your house. Your privacy has been violated; your home is your refuge and someone has intruded on your space. Now imagine they do more than just break in and steal some of your items, they also take the time to see what's in your fridge, rummage through your underwear drawer, read the various notes on your desk outlining a paper you've been working on, and more. Here I think we'd all find ourselves in agreement with Thomson's (1975: 299) assessment that, "the burglar's merely looking around in that way might make the episode feel worse than it otherwise would have done." Further, as I've noted in Basu (2021: 106), "there is something about another person attending to these private things about us that makes the rights violation indisputable." Further still, the wrongness of this additional looking around or prying or snooping behavior suggests that there are processes of coming to know things about others that constitute a violation of privacy and that we have a right against people coming to know things about us in those particular ways. 24 What the snooper violates is our contextual integrity and our ability to control, shape, and maintain our identity and how we present ourselves to others. Contextual integrity is a concept introduced by Nissenbaum (2004) to explain the importance of the control we exercise over how we present ourselves to different people in different contexts, something the snooper violates. As Nissenbaum (2004: 138) writes: norms of appropriateness dictate what information about persons is appropriate, or fitting, to reveal in a particular context. Generally, these norms circumscribe the type or nature of information about various individuals that, within a given context, is allowable, expected, or even demanded to be revealed. In medical contexts, it is appropriate to share details of our physical condition or, more specifically, the patient shares information about his or her physical condition with the physician but not vice versa; among friends we may pour over romantic entanglements (our own and those of others); to the bank or our creditors, we reveal financial information; with our professors, we discuss our own grades; at work, it is appropriate to discuss work-related goals and the details and quality of performance.
In this description of contextual integrity we see various forms of siloed forgetting. What we know about one another, what we attend to, and how we act varies from context to context. It would be inappropriate to email a stranger a picture of the rash on your thigh, but it would not be inappropriate to send such an email to your doctor. 23 As Nima Veiseh reports about his condition, "since forgetting is not really a luxury that I have, I have to genuinely learn to forgive, since forgiving and forgetting isn't something that, [it] isn't really the easiest thing for me." From https://www.wnycstudios.org/podcasts/takeaway/segments/what-if-you-could-remember-everything. Thanks to Alex Madva for pointing me to this source. 24 See Marmor (2015) for an argument along these lines.
It would, however, be inappropriate of your doctor, upon running into you at the store, to talk loudly about what they think of the rash on your thigh, even if their reason for doing so is to convey their medical advice to you. What your doctor violates in this case are important context boundaries. Information and topics of discussion that may be appropriate in one sphere are often inappropriate in others. To put the point simply, it is important to have boundaries, both in our physical world and, as I'll suggest with more cases, in our minds.

Civic Inattention
If I see you in a context in which you clearly did not wish to be seen, I ought to forget it. These practices of forgetting, or what Goffman (1963) calls civic inattention, are norms of etiquette that help ease our social functioning. 25 Returning to the case of the doctor, they're violating a rule of etiquette concerning how they should treat you in different contexts, both in their actions and their thoughts. To say that it's a matter of etiquette is not to dismiss the matter as morally unimportant. Many matters of etiquette are of deep moral significance. 26 It may not be morally important whether the salad plate is placed to the left of the forks (something I had to Google for the purposes of this example, so I hope I am correct), but it is morally important to us that people recognize us, that they do not ignore our presence, that they are polite to us, and that they are not rude to us. Rudeness betrays a kind of moral indifference.
These practices of etiquette, less like table settings and more like seating arrangements or not closing an elevator door on someone, concern the importance of mutual recognition, something central again to our practices of holding and letting go, something central to our practices of identity construction. Forgetting also plays an important role in our norms of etiquette. For example, it is common to forget embarrassing moments experienced by our friends. To silo off memories that people we care about wish for us to forget. Forgetting helps to ease our social interactions.

Intimacy
Our practices of forgetting are also central to intimacy. In a world where there are no barriers between contexts, we find ourselves in something akin to a Panopticon, i.e., we are always on display. In such a world we lose control over how we present ourselves, and in turn different kinds of relationships become more difficult to create and maintain. As a result, the self itself becomes more difficult to create and maintain. 27 This is a serious cost we would have to bear if we lost our ability to exercise a duty to forget. Lippert-Rasmussen (2017: 227) has suggested that this worry is overblown because as we lose our privacy it's still the case that intimacy can flourish through "the differential 25 For more on Goffman's work and how our practices of civic inattention ease social functioning, see Rini (2020). See also Rini and Cohen's (Forthcoming) discussion of panoptic gaslighting. As Rini and Cohen (Forthcoming: 11) note, deep fakes, i.e., highly realistic fake videos of a person, have the potential "to threaten memory and the existential bases of personhood." As a result, "a fabricated recording would be used to destabilize or oven overwrite first-personal, autobiographical memories." (Rini and Cohen Forthcoming: 11). acknowledgement of knowledge as common knowledge," and new norms of social interaction that would emerge as everyone's private lives became public. Some of our thoughts would certainly be less embarrassing if we knew that everyone had those same thoughts.
Notice, however, that for intimacy to flourish there still needs to be something kept private, something that is only ours, something that we control when it's shared. Further, there would still be some things that we'd ask people to forget about us to maintain intimacy. We, even on Lippert-Rasmussen's account, would develop ways of giving access to parts of ourselves to only some people and not others. Intimacy would be very different in this panopticon world with no privacy to our inner lives, but, some aspect of what makes intimacy intimatei.e., practices of remembering and forgettingwould remain. They'd just be reimagined. 28 2.6. Fruit of the Poisonous Tree Another response we have available to the snooper is that they acquired their beliefs in a morally suspect manner, and thus should be required to give them up. We can see this intuition captured in the fruit of the poisonous tree doctrine. In the US court system, even with its various faults, it is recognized that some evidence is inadmissible in court if it is fruit from the poisonous tree, i.e., we should ignore evidence if it was gathered in a way that violated a defendant's rights. 29 This doctrine captures an intuitive idea that it matters not just what evidence you have, but also how that evidence came into your possession. When it was evidence you didn't have a right to, you should forget it. 30 We can also find this intuitive thought throughout moral philosophy and epistemology more generally. That is, it matters not just that we believe what is true, but also that we get at what is true in the right way. We find this in the Kantian idea that it matters not just that we do the right thing, but we do the right thing for the right reason. We also find this in discussions of doxastic justification and the basing relation, i.e., it doesn't just matter that you believe p, but you also need to personally have good reasons for p. This thought is also present in discussions of knowledge, i.e., knowledge can't be the kind of thing we arrive at through luck, what you believe needs to be not only true but also connected to the truth in some further, non-luck-based way.

So, when should we forget?
Returning to contextual integrity and the fundamental interest we have in how we present ourselves to others, we can see this intuitive thought just discussed in §2.6 also in the claim that although normally we can't violate someone's right to privacy by believing something about them, we do when there is some privacy violation involved in how 28 As Alex Madva has suggested in his comments, in the dystopian future where technology prevents everything from being truly forgotten there will still be room for not "using" the wrong memory in the wrong context. That is, for developing mindfulness skills that resemble our current practices of forgetting, or at the very least perform the same function. 29 The doctrine was first established in 1920 in Silverthorne Lumber Co. v. United States, 251 U.S. 385. The phrase, "fruit of the poisonous tree," was first used in 1939 in Nardone v. United States, 308 U.S. 338. 30 It is also worth noting here that the reason this is an issue is because we are not perfectly rational creatures nor can we exercise perfect control over how what we remember affects and influences us. If we could remember something and not let it affect us, then there might not be a duty to forget. However, I'd counter that in the not letting it affect us we are treating the item as forgotten. you came to know this thing about them (see Marmor 2015). The intuitive thought echoed in these many examples extends also to forgetting: what you should forget is often a function of not just the content of the information, but also how you came to it in the first place. 31 As we see from these cases of our everyday practices of forgetting, forgetting facilitates another practice of significant moral value: our practice of identity construction. However, this duty to forget can be trumped by other considerations. As a result, this tentative duty to forget is best thought of as a prima facie duty. That is, it can be defeated. For a case with clear countervailing concerns, consider information that would be in the public's interest, e.g., cases of whistleblowing. 32 In such cases you may have obtained the information through the violation of a right to privacy, but that right to privacy is overridden by the public's right to know about the misconduct.

A New Threat to Forgetting
I now turn to how the account developed thus far can help us both recognize and address a heretofore neglected source of harm caused by technology and big data. The European Union's General Data Protection Regulation (GDPR) has introduced a right to be forgotten to combat the internet's permanence to our practices of identity construction. But there is a deeper threat to our identities, one concerning the ease of access to information, that my account of forgetting can bring to the surface.
Let us begin with some background on the right to be forgotten. The right to be forgotten has become enshrined in European law following the 2014 ruling in Google Spain v. AEPD and Mario Costeja González. The case centered on two newspaper announcements about a foreclosure auction on Mario Costeja González's home. As his social security debts had been resolved, Mario asked for the articles about the foreclosure to be removed as they were the first results that appeared when searching his name. That he owed debts in the past was no longer relevant, and like the cases we saw in §2, Mario didn't want his present to be defined by his past. The court determined that the articles could remain up, but that Google should no longer index them and had to remove the links to those articles. This case has given rise to the European Union's General Data Protection Regulation (GDPR) that gives individuals the right to have personal data erased, provided certain circumstances are met.
The valuable good that Mario Costeja González had lost and that the GDPR sought to restore is his contextual integrity. Google undermined Mario Costeja González's contextual integrity by removing his ability to present different aspects of himself to different people. Anyone can search Mario's name and discover this incident in his past, an incident that he does not want to define his future. Like juvenile arrest records following 31 For more on this approach see Munch (Ms). Munch similarly recognizes the importance and forgetting and doxastic wronging. In his paper, Munch (Ms) develops an account of why some beliefs, in particular those formed on the basis of injustice, ought to be disgorged, i.e., forgotten. Further, as Munch's view is motivated from principles concerning being the recipient of immorally acquired goods, there is also discussion of the fruit of the poisonous tree doctrine. See also Munch (2021). 32 For example, consider one of the positives of the threat I'll discuss in §3: the reach of the internet means that information can be shared widely and quickly. It is likely a desirable feature of the reach of the internet that videos of people being racist or cops killing unarmed civilians go viral. Without the internet these would be isolated incidents, but with the reach of the internet more people can see these everyday instances of racism and the systematic problems in policing. Thanks to Gabbrielle Johnson for helpfully bringing to my attention these alternative cases. one into adulthood, Mario's future has been constrained by what people will think of him on the basis of these search results.
How we appear in search results matters. Google has come to have a monopoly on how we get information. In the United States, 79.8% of all desktop searches originate from Google. In the UK it is 84.2%. In India, with a population of 1.38 billion, 94.19% of all desktop searches originate from Google. 33 Google has come to have this control over how we access information in large part due to the success of its search engine algorithm. Google's PageRank algorithm was inspired by academic citations where the number of citations a paper receives is taken as an indication of its importance. Larry Page and Sergey Brin extended this logic to hyperlinks on the web as a way of measuring the importance of webpages. The algorithm prioritizes sources that are linked from other sources and, as a result, the algorithm allowed Google to sort search results with well-established sources appearing higher in the results than random blogs. This feature distinguished Google from other search engine algorithms because Google searches would be able to give order to the sheer amount of information on the web. Google, unlike its competitors, could better sort search results according to importance. The success of Google's search engine algorithm has made 'Google' synonymous with the process of searching itself and a default authority accompanies the rankings delivered by Google. And as we've seen, Google has become the monolithic filter through which large populations have access to information.
Notice now how things would be different if the articles about Mario's home foreclosure were only available in print. The articles would still constrain his future and his ability to present himself as he wishes, but in a considerably smaller way. An article published in a local newspaper that was only available in print would function as forgotten after a time. Records may be kept at a library, but one would have to physically go to a library, request the records, and search through them by hand. Compare having to search through your physical letter correspondences versus the search function in your inbox. Searching the latter is considerably easier, making the emails more accessible and easily retrievable, and in some sense, more shallowly forgotten than the deeper kind of forgetting that happens for items that are harder to retrieve. Searching Google is considerably easier than hunting down a print copy of a newspaper.
To make matters worse for contextual integrity, Google will also push information to you that you hadn't intended to search for. If you were simply searching for Mario's email address, your top results would still include the articles about his home foreclosure and his social security debts. Digitized information about us archived on the internet has a far greater reach than ever before. The combined permanence of information and the new reach of that information means that it less is likely to be forgotten, let alone to function as forgotten, and this threatens our practices of identity shaping. The right to be forgotten is an attempt to build forgetting into the structure of the internet, but it cannot stop the harms from occurring.
For example, the internet facilitates a collapse of contextual integrity that is less common in the physical world. On this concept of context collapse, Frost-Arnold (2021: 437) writes: Context collapse is the blurring or merging of multiple contexts or audiences into one (boyd 2008;Marwick and boyd 2011). Internet theorist danah boyd is often 33 Statistics from https://www.statista.com/statistics/220534/googles-share-of-search-market-in-selectedcountries/. credited with first observing that context collapse is one of the key features of the internet by noting that "the lack of spatial, social, and temporal boundaries makes it difficult to maintain distinct social contexts" (boyd 2011, 49) […] Facebook users may intentionally collapse multiple contexts of social relations into one audience by sharing information with all their Facebook friends, which may include family members, close friends, acquaintances, and work colleagues.
Related to context collapse is the notion of context collision (see Davis and Jurgenson 2014) wherein what you thought you were sharing in one context, e.g., an Instagram story restricted to close friends that will disappear in 24 hours gets shared in a different context, e.g., someone screenshots your story and shares it on Reddit to millions of users. 34 Our current use of the internet has made a reasonable expectation of privacy almost impossible.
To demonstrate this, consider a more morally loaded case to bring out the privacy violation more vividly. Imagine that a jilted ex has made pornographic pictures of a job candidate available online. Although you had searched his name to find a copy of his latest publication, you may be shown those private images in addition to his PhilPapers page. This is a serious moral wrong that you could find yourself simply stumbling into through no fault of your own. To illustrate the mental burden that accompanies these more egregious forms of context collapse, consider the following quote that Noble (2018: 120) presents from an interview conducted with an anonymous victim of the revenge porn website isanyoneup.com: I was submitted to isanyoneup.com by my ex-boyfriend. I am confronted by friends, family and strangers that they have seen me naked online everyday. … You may think it's funny but sometimes [I] don't want to leave my house and go to the mall with my family because I fear somebody will come up to me while I'm with my mother and mention it. My sisters … are ashamed to be related with me and want to lie to their friends that they are my sisters. I am a disgrace to my family. … My self worth has gone out the window and I worry I may never get it back. This keeps me one step away from happiness every single day. I don't know what to do anymore.
If you searched someone's name and discovered leaked nude pictures, presumably you should forget what you've just seen. The mere act of scrolling through the pictures, let alone were you to also download them and email them to others, would be a clear violation of their privacy. The person whose pictures those are never intended to have them on the internet for the world to see, and they have a right against them being available for you to so easily find. Although it shouldn't matter to your evaluation of a job candidate that these pictures exist online, the reason revenge porn is revenge is because leaked nude photographs do in fact affect people's opinion of others.
To make matters worse, search engine algorithms are able to not only access and reveal information about us, but they also have a power to construct and dictate intimate parts of our lives. Consider a case I discuss elsewhere, that of Janet Kim, from the hit Canadian sitcom Kim's Convenience. 35 Janet, upon trying to make a website to 34 Alex Madva has told me that if I didn't mention George Costanza howling about "worlds colliding" then he would never forgive me, let alone forget. I hope this footnote suffices. 35 Much of this section on Google searches and identity is drawn from my post "The Internet Never Forgets: How Google Shapes and Cements Our Identities" published here: https://blogs.cardiff.ac. uk/openfordebate/the-internet-never-forgets-how-google-shapes-and-cements-our-identities/.
showcase her photography, wishes her parents had given her a more unique name so her website would appear higher in the search results. Our revenge porn victim wishes to disappear, but the internet makes that incredibly difficult. Janet on the other hand wishes to stand out, but the sheer volume of information on the internet also makes that incredibly difficult. There is little control that we as individuals can exercise here with regard to how we are presented online. There is little control we can exercise over our identity shaping practices.
Notice that what matters here is not only how the information is curated, e.g., what appears as the top results when searching for your name, identity, occupation, etc., but also what matters is the availability and the permanence of information online and how that undermines the goods that forgetting seeks to facilitate. The former, the topic of curating search results, has received much attention, notably in Noble's (2018) discussion of how search engines can reinforce systems of oppression. The latter point, about the combined threat of the permanence and reach of the internet, has not received as much attention. Changing our norms of information curation (as through the right to be forgotten) will help somewhat, but there's a deeper issue about access and permanence. At least the burglar has to put in some work to discover private information about us.
Where we can exercise some control is in how we respond to information we discover about others on the internet. That is, we can update our norms governing how people come to know things about us, what they store in their memory about us, and how it is stored, e.g., how accessible it is, in order to enable the valuable forms of forgetting that the internet threatens. To put it bluntly, what we need is a norm of the following kind: when we learn something about another due to an act that violated another's privacy, unless there are countervailing concerns of greater moral weight, we are required to forget what we've learned. That is, we have a prima facie duty to forget what we learn in ways that undermine another's right to privacy.

Two Versions of the Control Worry
One challenge to a doxastic wronging view of forgetting, familiar to doxastic wronging views about belief, is that we can't directly control our mental processes, such as what we forget or believe. This challenge arises for any account that suggests we ought to structure our attitudes or have attitudes of a particular kind. It follows from the principle that we cannot control what we forget, and from a general commitment to ought implies can, that we cannot be held responsible for things that are not within our control. Our attitudes are generally taken to be a paradigm instance of something not in our control, and thus something for which we lack responsibility. For example, I cannot, at this very moment, believe at will the counterfactual that the lights in my office are on. Similarly, I cannot will myself to believe that it's raining when I can see directly that it is not. If we lack control over what we believe, and we're only obligated to do what's within our control, then we can't be obligated to either believe or not believe.
Extending this worry to the case of forgetting, the challenge becomes a question of whether we have sufficient control over the forms of forgetting such that we could be obligated to forget. It would certainly be better to forget that you've seen someone's leaked nude photos, but if it's impossible to forget, you can't be required to do so.
In general, I take the control worry to be a red herring, i.e., distracting from what is really at issue. The issue being that we do in fact hold people accountable for what they believe, what they forget, and generally how they structure their attitudes. That ought implies can is philosophical orthodoxy obscures that our everyday practices speak strongly against it. Furthermore, the issue is a distracting red herring because regardless of what position you take with regard to control and responsibility, there is at least one argument for why we can nonetheless be held responsible.
In earlier work (see Basu 2019b) I've demonstrated that no matter your position with regard to the argument that deontic notions such as ought and should and moral notions such as praise and blame do not apply to mental states outside of our control, there is a way available to establish that ought, should, and moral notions more generally apply to our mental states. Here is the standard argument for the control worry as presented in my earlier paper: 1. For deontological concepts such as obligation or duty to apply in the domain of epistemology, agents need to have voluntary control over their beliefs. 2. It is not the case that we have sufficient voluntary control over our beliefs for deontological concepts such as obligation or duty to apply in the domain of epistemology. 3. Therefore, deontological concepts such as obligation or duty do not apply in the domain of epistemology.
There are two paths of resistance available against this argument. The first is to deny premise one. That is, to argue that we can be held responsible and blamed for actions and intentions that are unintentional or involuntary in the same way as beliefs are claimed to be (see Hieronymi 2006Hieronymi , 2008. To extend that to forgetting, even if we didn't have any kind of control over forgetting we can still be held responsible and blamed for what we forget just as we can be held responsible and blamed for other things not under our control. The second path is to deny premise two. First, one could argue that we have direct control over our attitudes. For such an approach see Bernecker (2018) for a survey of psychological studies demonstrating that we have control over forgetting, that is, there are non-drastic techniques available for forgetting specific things, i.e., directed forgetting. That is, participants instructed to forget some information, e.g., a list of words, are able to do so. Alternatively, one could argue that whatever the sufficient voluntary control is that is required for the notions to apply, we have it (Steup 2000;Shah 2002;Weatherson 2008;Basu and Schroeder 2019). 36 I personally favor this approach. The cases involving forgetting surveyed thus far all demonstrate some level of control, albeit indirect control, over what we forget in the same way we have some indirect control over what we believe. We cannot believe at will and some beliefs may come to us unwillingly, but we can guide how we believe. That is, we can train ourselves to be less quick to believe, it is up to us when we end inquiry and settle on a belief, etc. Similarly, although we cannot forget at will, if we don't tend to our memories they can often fade away even if it's the case that some memories can't be dislodged no matter what we try. Recall my point that forgetting is a process. There is some control we are able to exercise, albeit usually indirectly, and that is sufficient for blame and responsibility to apply.
However, there is a harder challenge lurking within this control worry. Perhaps the fundamental interests that we aim to protect through a right to privacy are impossible given the world we find ourselves in, a world in which our ability to forget is being eroded by companies whose raison d'être is to preserve a record of everything so 36 See also Smith (2005) for why we should reject the volitional view of responsibility altogether. See also Singer et al. (2021) for an argument for how forgetting is reasons-responsive. those records can be sold to the highest bidder. That is, the more information about us that is preserved, the more information the internet makes easily accessible about us, the more central role these large caches of information play in our lives, the harder it will be to maintain any meaningful sense of forgetting. The real challenge to forgetting is not our own volitional abilities, but the challenge that the combined publicity and permanence of the internet age pose to our ability to forget. This challenge, I fear, is not so easily answered. To answer this challenge we must change the world. 37

Conclusion
In this paper I have argued that morality bears on what we should forget. I have not been able to address every aspect of forgetting but I have argued that just as there are moral demands on believing, so too there are moral demands on forgetting. 38 Forgetting is an integral part of our moral duties to others because some aspects of our identity are meant to be siloed or impermanent. Drawing on Lindemann's (2016) account of holding and letting go, I have argued that our practices of identity shaping depend on our ability to grow and change over time. Notice that this isn't the same as an unmitigated gain of information about someone, it's recognizing that they are no longer the same person as who they were before. The distinctive harm that the permanence of some content has for us is what prompts the moral duty to forget. However, the digital age poses a new challenge to forgetting.
As I've noted, the moral duty to forget arises in response to the detrimental effects of the permanence of information about us. The new threat of the digital age not only exacerbates this detriment, but also it introduces a new feature that makes the detriment worse than before. It is not only the case that information about us will be preserved forever online, but also that information is easily accessible. These are two distinct threats to our identity shaping practices. The first threat we can address through our practices of forgetting. Forgetting is something we do over time, it is a practice or a habit that can be cultivated in better or worse ways. And we are responsible for how we engage in this practice, or whether we engage at all. The second threat, however, should serve as a call to action. We are losing valuable goods enabled by forgetting. Addressing the second challenge cannot be done individually. Addressing the second challenge requires changing our processes for accessing information about others online. Addressing the second challenge means fundamentally changing our relationship to search technology. Addressing the second challenge requires changing the internet to make sure the internet can forget. 39 37 As much as we'd love more direction than "change the world," note the similarity between this result and Johnson's (2020aJohnson's ( , 2020b answer to the problem of algorithmic bias. As long as injustice exists, algorithms trained on an unjust world will replicate those biases. If we want to get rid of these biases, we have to change the world. 38 For more on forgetting see Bernecker (2018) for the degree to which we can be held responsible for both intentional and unintentional forgettings. See also Murray et al. (2019) and Irving et al. (Forthcoming) for more on the responsibility question, and for whether forgetting is irrational see Broome (2013) or for whether forgetting is simply unfortunate see Williamson (2000). See also Frise (2018) and Rieff (2016). 39 Thanks to audiences at Nanyang Technological University, Cal State Long Beach, USC, DEX VII at UC Davis, UC Riverside, and a special thanks to the audience at the 16th Episteme Conference where I first presented this paper, especially to my commentator Jason Kawall. Thanks also to Maegan Fairchild and