Content-Era Ethics

New media forms affect a culture, in part, by reshaping what is seeable and sayable: what “ideas,” as Neil Postman once put it, “we can conveniently express.” In this essay, I ask what one of today’s major new media forms—viral, digital “content”—compels us to see and say. To address that question, I em-brace a makeshift, hybrid methodology, informed by theory, sociology, arts criticism, and the digital humanities, and eschewing media theoretical orthodoxies that have been dominant across the humanities (namely: an exaggerated emphasis on the “medium” at the expense of the “message”). From this poly-glot perspective, I analyze content contained in a database that I have compiled, indexing 205,147 of the most-shared pieces of viral media on sites like Facebook and Twitter, from 2014 to 2019. After survey-ing this content’s basic features, I focus on one, particularly popular and quintessential content genre, which I call the “uplifting anecdote”: a short, sentimental account of a heroic act. The uplifting anec-dote, I argue, promotes a novel type of ethics, ideally suited to the content economy. I then track this ethics’ dissemination into the broader culture, through a discussion of two prominent, aesthetic artifacts: George Saunders’ prize-winning, best-selling novel, Lincoln in the Bardo (2017), and NBC’s popular sitcom, The Good Place (2016-2020).

themselves." 4 For Jia Tolentino, such media "brings the 'I' into everything": "it's as if we've been placed on a lookout that oversees the entire world and given a pair of binoculars that makes everything look like our own reflection." 5 For Wendy Chun, it renders YouTube's "You" digital life's definitive pronoun. 6 In either case, the implication is the same. Each piece of content approaches the user, not merely as a source of entertainment, but also as a fashion statement: "What would this look like on me?" New media forms, Neil Postman once argued, affect a culture by reshaping what is seeable and sayable, or "what ideas we can conveniently express." 7 What does content compel us to see and say? In this article, I approach that question from a perspective somewhere between the disciplinary coordinates of media theory, sociological internet studies, literary criticism and the digital humanities. Like a theorist, I approach a broad, cultural question by means of abstract thought and arts analysis. Unlike a (typical) media theorist, however, I focus on the object against which media theory is defined: content itself. Rather than simply examine the institutional, legal, or technological structures that shape social media entertainment, I also analyze that entertainment's look and feel. I do not, however, do so in precisely the manner that a more sociological practitioner of media studies might: by means of a largely thematic "content analysis." Rather, I do so more in the mode of a literary critic, or digital humanist working within an arts discipline. I attend not only to the content's thematics, but also to its aesthetic, narrative, and generic elements, as revealed by "close" and "distant" reading. The convictions behind this approach are simple: that the meanings encoded in a medium are very real contributors to its cultural effect; and that the aesthetic forms through which those meanings are conveyed make a difference, too. In sum, my approach is pluralist. For that very reason, I will not dwell much longer on questions of method. A lengthy discussion of the pros and cons of McLuhanism or "distant reading" would leave little room to address the question at hand. 8 To address that question, I will examine content indexed in a database that I have compiled using public and proprietary sources and tools (Newswhip and Buzzsumo). 9 The database contains a list of 205,147 pieces of content that have, throughout the past five years (2014-2019) gone "viral." By that, I mean that they have earned 500,000 or more shares on popular social networking sites, particularly Facebook and Twitter. This viral content includes much of the most popular content produced by top producers, like NYTimes.com and Buzzfeed.com; it also includes a critical mass of mega-viral content produced by more obscure outlets. Here, I will focus, primarily, on one segment of the data in the database: the headline text. 10 To analyze that text, I will use close and distant methods (Topic Modeling, Sentiment Analysis, Naïve Bayes Classification). To minimize metadiscussion, however, I will move in-depth descriptions of procedure to endnotes.
In addition to analyzing content en masse, I will focus more narrowly on one particularly popular and quintessential content genre, which I call the "uplifting anecdote." Uplifting anecdotes are short, heartwarming stories (or groups of stories), presented as true, like Reshareworthy's "Incredible Dog Who Saved the Life of Newborn Baby Has Lasting Legacy" or Bored Panda's "30 Suicide Survivors Share How Happy They Finally Are." 11 Generally, they are sentimental, in the pejorative, post-1900 sense of the word. More particularly, they evoke an emotion thatas Jill Abramson has quipped -has no name: "the feeling of having one's faith in humanity restored" (In this respect, they fall into the broader category of what sociologists call "eudaimonic media"). 12 Uplifting anecdotes are pervasive, comprising an estimated one-ninth, or 11.4 percent, of the content listed in my archive. 13 They are also representative. As I will show, they embody many of content's more common features. One of those features is an overwhelming concern with the ethical or moral. By focusing on the uplifting anecdote, I examine how social media content's rise to prominence is reshaping popular ethics.
Mainly, I argue that social media content, thanks to its particular pursuit of "prosumption," abets a simultaneous expansion and thinning of ethics. It promotes heightened moral concern, but simultaneously reduces moral action to aesthetic self-expressionnamely, the self-curation involved in acts like "sharing," "commenting," and "liking." Whereas mass media may have, as Benjamin argued, enabled an "aestheticization of politics," content contributes to a parallel and particular aestheticization of ethics. 14 My analysis will proceed in three parts. First, I will look broadly at contentits key features, and their ties to prosumption. Next, I will look more closely at the uplifting anecdote, conceived of as an embodiment of those features, and analyze the ethics that it promotes. Finally, to unpack the implications of that ethics, I will trace some of its dissemination into the broader culture.
In particular, I will reflect upon the uplifting anecdote and its ethics' influence in relation to two cultural artifacts: NBC's sitcom The Good Place and George Saunders's novel Lincoln in the Bardo. Emerging at around the same time , the TV show and novel both take place in afterlife spaces -the "Good Place" and "Bardo"that, as I will argue, operate as all-too-apt allegories for social media, conveying its pervasive atmosphere of moral and ethical scrutiny. The show and novel also reflect, more specifically, on social media content's aestheticized ethics. Both, as I will demonstrate, were not only loosely inspired by uplifting anecdotes, but also channel the genre in their content and form. In the process, they embrace divergent responses to a world of ever-present and yet etiolated ethics.

Content, Two Ways
Two major schemata describe the content in my archive particularly well, capturing its major thematic, generic, and aesthetic features. The first schema orients the content between two poles: the domestic and heartwarming and the political and corrosive. The second schema divides the content into three major categories: content concerning (1) real, sensational events, (2) the self, or (3) artifice or craft. Close and distant reading, taken together, suggest those schemata.
To begin with the first schema, a critical mass of content approaches (if without always fully embodying) one of two opposing poles. The first pole represents content that is sentimental, uplifting, or unifying. This content often focuses on ethical, private, or domestic realms. Its topics are family life or other classic objects of sentimental concern: the vulnerable among us (as Phil Fisher aptly defines them), like animals, children, or the elderly. 15 Examples include The Dodo's "Man Builds 'Dog Train' To Take Rescued Pups Out On Little Adventures" or Bitecharge's "I am a Faithful Wife. What Type of Wife Are You?" 16 The second pole represents content that is corrosive, upsetting, or divisive. This content often focuses on political or public realms. In recent years, it has mostly fixated on Donald Trump. Examples include Shoebat.com's "Hillary Clinton: Christians in America Must Deny Their Faith in Christianity" or The Independent's "Donald Trump May be the Dimmest President the U.S. has Ever Had." 17 Multiple methods suggest those two tendencies' ubiquity, as well as the slight dominance of the first. A topic model, for example, divides the headline text of the content into 150 (strikingly straightforward) topics. The model was run on 205,147 documentsthe headlinescomprising 2,399,270 words. 18 The top three topics, ordered by relative weight, can be titled: "Family," "Animal Rescue" (one subgenre of the uplifting anecdote), and "Trump and the Economy" (Table 1). This trio suggests the utility of the two-sided schema, with the first two most abundant topics suggesting the sentimental or domestic and the third suggesting the political or corrosive. A closer look at the topics confirms this schema's utility. Of the 150 total topics, almost half (68) fall cleanly into one of the two major categories, with sentimental or domestic topics (like "parenting," "childbirth" or "DIY Home") possessing a cumulative weight of 0.311, and political or corrosive topics (like "Russian Investigation," "Trump vs. Clinton," and "Police Brutality/BLM") possessing a cumulative weight of 0.301. Many of the rest of the topics have at least some affinity with one category or another. Otherwise, they are diverse (they include, for example, "science facts," "restaurant deals," and "amazing photos"). More simple analyses, like word frequencies, reinforce this same schema. A plurality of the fifteen most frequently appearing nouns in the headlines, for example, conjure home and hearth ("house," "home," "family," "dog," "school," "baby"), while the single most frequent, "Trump," invokes political controversy (though its use does not outnumber that of the domestic/sentimental words). Other analyses yield similar results. A sentiment analysis of the headlines, using Saif Mohammad's NRC Lexicon (which associates English words with eight different feelings), is consistent with the slight dominance of the domestic and unifying over the political and controversial. 19 It shows that, relative to all of the words in the lexicon, the words in the content headlines embody a higher ratio of positive and prosocial feelings -like "joy" or "trust"to negative or corrosive feelingslike "anger," "fear," or "disgust" (Figure 1).
A second, useful schema is tripartite, dividing the content into three major categories. First, there is content that describes real, sensational events. This includes, for example, the genres of the uplifting anecdote and the upsetting anecdote (its less abundant foil). Second, there is content that indulges its audiences in navel gazing. This includes, for example, the genres of the quiz (selfdescription) or how-to guide (self-improvement). Third, there is content that is aesthetic or fictitious in nature or concern, rather than (allegedly) informative. This includes, for example, the genres of the music video or craft-related content (e.g., "Seven breathtaking sandcastles"). Most content appears to belong to at least one of these three categories. Close inspection of multiple samples of the content confirms that rough impression. Almost all of the first thirty consecutive pieces of content listed in my archive (Table 2), for examplealso the thirty most shared can be sorted into the three proposed categories. Category one claims the political news, uplifting anecdotes, and celebrity deaths; category two claims the quizzes and self-celebratory think-pieces; category three claims the music videos and photography. Computation reinforces this schema's comprehensiveness. Consider again the results of the topic model, a list of 150 major topics ranked by weight (Table 2). Category one claims two of the three most abundant topics, "Animal Rescue" and "Trump and the Economy," as well as fifty-nine other lower ranking or redundant topics, like "Russia Investigation," "Traffic Accidents," and "Animal and Human Rescues." Category two then claims eighteen of the remaining topics, including "Weight Loss," "Hair and Makeup," and "Personality Type Quiz." Category three encompasses another thirty-three topics, including "Music Videos," "Bollywood," and "Tattoos and Arts and Crafts." All in all, more than 2/3 of the 150 topics fall into these three categories. This analysis suggests, moreover, that the first category of content, about sensational events, dominates the corpus: topic groupings associated with it contain the largest number of individual topics (61 of 150) 20 and have a dominant cumulative weight (0.52698). 21 Such content is also, it is worth noting, largely divided in accordance with the first schema: between the sentimental, uplifting, or domestic (e.g., uplifting anecdotes) and the political, corrosive, or controversial (e.g., upsetting anecdotes).
Headline URL  In sum, viral social media content has dominant features. It veers toward either the sentimental and ethical or the corrosive and political, and it satisfies desires for sensation, self-actualization, and aesthetic delectation. Many of those features' dominance, moreover, can be seen as a product of the content's more basic distinguishing quality: its pursuit of prosumption. I'll discuss only three examples, relevant to my subsequent discussion of the uplifting anecdote. First, social media content's particular tendency toward sentiment andby extensionethics, is fitting. Content and sentiment, in general, go well together. As per its pejorative definition, content is utilitarian, designed to achieve economic aims. Sentiment is commonly defined as "didactic rather than mimetic," designed to persuade more than accurately depict. 22 Social media content, in particular, still more appropriately lends itself to the sentimental. Indeed, the aesthetic is ideally suited to inspire prosumption. It stokes the same two desires for self-expression and connection that compel shares, comments, and likes. When we react to sentimental media (as opposed to, say, pornography), we want to publicly own our feelingsself-congratulation, as many have noted, is central to sentiment's reception. Moreover, we are liable to want to commune with otherssentiment's standard byproduct, as Joanna Dobson has put it, is a "desire for bonding." 23 Unsurprising, then, not only that social media content is so often sentimental, but also that the platforms and hubs that circulate it embrace the communal aesthetic. Facebook creates "friendiversary" videos chronicling past digital interactions between users; Buzzfeed and Bored Panda exhort content creators to "join the . . . Community," without promise of employment or pay. 24 In place of monetary compensation, they offer a saccharine ideal of amateur creativity: "What I love about Bored Panda," one contributor writes in a promotional blog post, "is that it's almost entirely usergenerated . . . It's so much more emotionally engaging than hearing about something from a journalist." 25 Other major features of content follow clearly from its core pursuits, namely, its focus on the self and on artifice or craft. That a type of entertainment designed to operate as part of its audiences' public image would emphasize the self is true almost by tautology. Quizzes are so popular on social media because their results can be shared. Prosumption's pursuit also motivates an emphasis on artifice. Content relies, for its profits, on audiences' engagement in acts of self-expressive making. Naturally, then, it glorifies the activity. The link between idealizing amateur creation and attracting digital responses need not be subtle. Videos that depict influencers baking bread or braiding hair will often directly ask viewers to respond by posting their own attempts. 26 Bored Panda's brand, meanwhile, is not only sentimental, but also sentimental about do-it-yourself or nonprofessional creativity ("entirely user-generated").
The genre that I call the "uplifting anecdote" is not merely abundant, but representative, as the two major schemata under discussion suggest. Consider its relationship to each schema. The uplifting anecdote keys directly into what the first schema tells us is content's major tendency: toward the sentimental, domestic, or heartwarming. The content genre concerns a sentimentalized entity (like a puppy or mother) and tells a saccharine story that ends happily. The uplifting anecdote genre also keys into all three of the categories of content under the second schema, satisfying demands for sensation, self-improvement, and aesthetic experience. Its affinity with the first should be obviousit always tells an allegedly true, remarkable story. Its affinities with the second and third concerning self and craft are less so. In this essay's next section, I will establish those less conspicuous affinities and discuss their consequences for the particular type of ethics that the content genre promotes.

Ethics and the Uplifting Anecdote
The uplifting anecdote is not a wholly novel form of popular entertainment. Generally, it resembles a large body of prior sentimental media, from Uncle Tom's Cabin to the sitcom's "very special episode." More particularly, it is preempted by a smaller collection of media that is both sentimental and anecdotal, like the saccharine TV advertisement (" . . . Priceless") or political speech's call-out segment ("Let me tell you about my good friend Jim, from Georgia . . . "). Contributors to the "Viral Texts Project" at Northeastern have discovered what may be the uplifting anecdote's clearest ancestor. Setting out to discover the nineteenth-century version of "viral texts"or texts that spread from newspaper to newspaperthey identified the genre that they call the "vignette": a short, true story that is often, like the uplifting anecdote, saccharine or heartwarming. One popular nineteenth-century vignette, for example, tells the story of a widower's discovery of a deathbed letter from his late wife. 27 Like most of its precursors, the uplifting anecdote concerns itself with ethics. In general, sentimental media invoke the ethical in one of two ways (there are others). They aim to inspire sympathetic or altruistic feelings toward the vulnerable parties on whom, as Fisher notes, they often focus. Or they invite audiences to take pleasure in the depiction of those feelings, particularly in the context of familial or proto-familial relations (between mothers and children, friends who are like sisters, etc.). Sometimes, such responses are presumed -Dickens assumes his readers' capacity to sympathize with dying children. Sometimes, they are compelled -Beecher Stowe attempts to awaken her readers' latent compassion for enslaved Americans. The uplifting anecdote invokes ethics in a similar way. In general, it relies on audiences' feelings of sympathy for diminutive figuresdogs, infants, children -who are in peril. It relies, also, on audiences' feelings of moral approval when some heroic person or animal saves the day. More often than not, such feelings are assumed. The genre mostly asks audiences to feel quasiuniversal, uncontroversial forms of sympathy, for parentless puppies, harried mothers, or varied types of ugly ducklings. In a much smaller number of cases, they compel more divisive responses, asking audiences to feel in ways that code politically right or left. An uplifting anecdote from Breitbart, for example, asks audiences to celebrate a gun owner's heroism ("Concealed Permit Holder Stops Attempted Mass Shooting in Chicago"). An uplifting anecdote from Buzzfeed asks audiences to sentimentalize the struggle for LGBTQ rights ("This Gay Couple Re-Created their Pride Parade Photo Seven Years Later and it Has People Emotional"). 28 Nevertheless, uplifting anecdotes specifically, like content genres generally, differ from their print or television precursors. They do so on the same grounds, in that they profit immediately, and directly, by inspiring "prosumption." For this reason, they promulgate a particular, and arguably problematic, variety of ethics. In the subsequent paragraphs, I will make this point by drawing not only on close analysis of uplifting anecdotes, but also on the results of a final digital experiment. 29 For this experiment, I trained a Naïve Bayes classifier on a handlabeled dataset of the headlines of 1022 uplifting anecdotes and of 1022 other pieces of content. The classifier, trained using ten-fold cross-validation, could distinguish between the two groups of headlines with 76.7 percent accuracy, on the basis of some trends (this rate surpasses that achieved by randomized labelingor the "null baseline"at 51.03 percent; it is also comparable to rates achieved by similar experiments). 30 A sample of these trends are depicted in Table 3, showing the five words most and least likely to indicate uplifting anecdote headlines, as expressed by a ratio (the number of appearances in uplifting anecdote headlines over the number of appearances in other headlines, adjusted to prevent dividing by zero). Many of these results simply confirm the uplifting anecdote's nature as I have defined it, and on the basis of which the tagging was done. They show, for example, that uplifting anecdote headlines are more likely than those of other content to refer to sentimentalized beings, particularly children and dogs. They also show that uplifting anecdotes are, by and large, at odds with the major classes of politically divisive content. Three of the four words least predictive of uplifting anecdote headlines were "Trump," "Donald," and "President" (though those results are influenced by the abundance of those words). Other of the classifier's results, however, are more unexpected. Those   In general, where critics have discussed the ethics of sentimental media, they have tended to do oneor bothof two things. Where they have praised sentimental media's ethics, they have focused on its sincere pursuit of moral persuasion: its ability to make audiences sympathize with, and potentially assist, members of vulnerable groups. Fisher, for example, has argued that though sentimental novels like Uncle Tom's Cabin may engage in forms of stereotyping that seem distasteful in retrospect, such crude depictions are a necessary first step toward humane consideration. 31 Where critics have critiqued sentimental media's ethics, they have focused on its economic aims: how, in pursuit of profit, it flatters its audiences into narcissistic quietism. That is Ann Douglas's argument, which condemns the sentimental as a consumerist aesthetic, designed to indulge audiences in the shallow, self-satisfied feelings. James Baldwin famously describes those feelings as the "ostentatious parading of excessive and spurious emotion." 32 Put another way, critics' positive or negative evaluations of sentimentalism have often hinged on how completely they conceive of it as conforming to a definition of content in the pejorative sense. Unsurprising, then, that the uplifting anecdote merits something less like the first mode of critical celebration and something more like the second mode of critical censure.
It does so, though, in its own particular manner. Thanks to its pursuit of prosumption, it promotes a shallow, inert sort of ethics, equating moral action with aesthetic self-expression. It does so, first, by means of its aforementioned emphasis on the self (as per the second schema). Uplifting anecdotes differ from some if not all other sentimental media by virtue of one of their key features: their happy endings. Rather than focus primarily on sufferinginviting audiences to mourn they focus more on the heroic acts through which that suffering is alleviatedinviting audiences to celebrate. They then equate those heroic acts with the acts of emotive self-expression through the sharing, liking, and commenting that they attempt to inspire. Consider, for example, the following comparison, between one classic type of sentimental mediaa Save the Children Adand a particularly viral, uplifting anecdote: "Mexican Bakers Make Pan Dulce For Hundreds of Harvey Victims After Being Trapped By Floods" (Figure 2). 33 The primary and most overt aim of the Save the Children ad (to oversimplify only slightly) is to encourage charitable contributions -for this reason, it dwells on the child's suffering. Indeed, this is the approach that most GoFundMes take. The main aim of the uplifting anecdote, however, is to compel not only consumption, but also prosumption. For this reason, the anecdote focuses less on the bakers' own heroic act than it does on the positive response that the bakers' heroism is inspiring, both on and offline. Rather than describe or interview the bakers themselves, the anecdote's text moves from quoting their kvelling manager to recapitulating social media responses to their feat: "thousands of people reacted to the bakers' 'heartwarming' feat on social media," the article's text reports, before quoting gushing comments at length. 34 That text flanks an image, not of the bakers, but of a Facebook post about the event, featuring their image. The uplifting event being narrated, here, is less the bakers' generosity than its cyber-celebration, in which the viewer can also now participate.
The classifier provides additional evidence for this idea, in one of the more interesting trends that it reveals: relative to non-uplifting anecdote headlines, uplifting anecdote headlines are distinguished by the frequency with which they refer, in meta-fashion, to the very processes of content circulation. They are 5.5 times more likely to use the word "internet" (as in "the internet is going crazy over . . ", " . . . broke the internet"); 4 times more likely to use the word "viral" ("This viral story . . . "); and 5 to 10 times more likely to use various words that preemptively describe the content's emotional impact ("touching," "adorable," "sweet," "wow," "heartwarming," "powerful," etc.). This abundance of metadiscussion is very likely the result of their insistence on an equation between ethics and online prosumption.
Indeed, uplifting anecdotes tend to equate moral action not only with selfexpression, but also with the amateur acts of aesthetic creation and curation that prosumption involves. (Here we see the genre's emphasis on artifice or craft.) The classifier provides an initial indication of this trend. It suggests that uplifting anecdote headlines, even more than those of other content, use certain words to connote aesthetics or making. They're particularly likely, for example, to use the words "beauty" (7 times more likely) or "beautiful" (3.7), to use words that call attention to their visual medium like "pictures" (8) or "photo" (3.8), or to use other words that suggest acts of making or their results, "sing" (5.5), "write" (4), "written" (4), etc. Whether uplifting anecdotes refer to aesthetics or making more than other content, in any case, is not as important as the fact that they do so excessively. Indeed, they enfold the idea of artifice into their ethical system in two ways. First, they frequently describe acts of moral heroism that are also acts of amateur creation. Here are but a small sample of anecdotes that do so: a man takes photos of himself in a pink tutu to cheer up his sick wife; a woman draws pictures celebrating her boyfriend's love; a man makes cartoons celebrating his relationship with his girlfriend; a gay couple creates photo montages of themselves at Pride; a man creates photo montages of his dog and cat who are best friends; a man splices together Harry Potter clips, featuring the character Severus Snape, making people "feel things"; a child signs a Christmas concert in ESL for her deaf parents; a man signs a wedding song in ESL for his (not deaf) daughter; a mother invents a harness for disabled children; a boy invents a device to cool down hot cars after he learns of one killing an infant; and so on. 35 Second, they equate moral heroism not only with creation, but also with curation: the recognition and dissemination of aesthetically pleasing or beautiful entities that might be involved in sharing, liking, or commenting on content. Here, the uplifting anecdote enacts its most dramatic elision of the ethical and aesthetic. As an example, consider one of the most abundant subgenres of uplifting anecdote. In this subgenre, some human or animal not previously regarded to be beautiful or desirable is acknowledged, recognized, or depictedstirringly, heroicallyas such. The content may subvert invidious stereotypes regarding what is beautiful or attractive, but may also inadvertently affirm them by assuming they are shared. Examples include "Girl Mistakes Bride for Real Life Princess From Book She's Holding and the Reaction Melts Everyone's Hearts" (the bride, unlike the princess in the girl's book, is black), "This Woman Was Nervous About Her Photoshoot With Fiancé, But The Results Won the Internet" (the woman is overweight, and the photographer celebrated rather than hid that fact), "Teen With Down Syndrome is determined to become a Model" (she succeeds), "Teen Bullied for Her Incredibly Dark Skin Color Becomes a Model, Takes the Internet By Storm," and "Rare Kitten Born With Two Faces Grows Into the Most Beautiful Cat Ever." 36 Audiences can enact the aesthetic recognition that those stories praise in recursive chains of shares, comments, and likes. The examples reveal an increasingly pervasive attitude that to have one's human worth recognized is to be deemed an appropriate candidate for sexual objectification. That puzzling millennial ideology has gained traction among so-called "incels" and members of other erotic online subcultures. The analysis also supports an armchair observation that social media is in part responsible for the post-circa 2010 fixation on the ethics of representation in pop culture, as much if not more than in other realms (politics, professions). I am thinking here of #OscarsSoWhite or the celebration of the "strong female protagonist." A medium bent on compelling users to engage in representational acts, unsurprisingly, invests the symbolic realm with supreme moral import.
Of course, not all social media users embrace the uplifting anecdote's moral implications. Even the dissent that the genre inspires, however, is often colored by its ethics. To be sure, some reject the genre from a position beyond the pale of its presumptions. They tend to vocalize a classic critique, which might apply to any type of sentimental media: that emphases on personal altruism or sympathetic feeling can serve as narcissistic distractions from the realer, more pressing problems of politics. A recent uplifting anecdote, for example, which described a CEO's $22,000 contribution to the relief of Philadelphia students' school lunch debts, inspired this type of response. Celebrating the CEO's generosity, some argued, distracted from the deeper issue of the Republican party's anti-welfare policies opposing free lunch programs and debt relief. 37 Others, however, take issue with uplifting anecdotes from perspectives that accept more ofindeed, are inspired bytheir particular premises. Consider, for example, one logical consequence of the uplifting anecdote's consistent suggestion that all moral action

C O N T E N T -E R A E T H I C S
310 occurs on the plane of symbolic self-expression. If this is the case, then the predominant ethical crime is no longer either inaction or incorrect action: rather, it is false self-depiction. Indeed, much of the ethical debate that goes on, in the realm of social media content, concerns how sincere, accurate, or authentic a person's performance appears to be. Such authenticity is at stake in the classic complaint of the internet nihilist or "troll," who objects to all inevitably self-serving "virtue signaling." It is also at stake in the objection that users tend to levy (if any) at otherwise uncontroversial uplifting anecdotes (indeed, this type of response is more typical than that which the school lunch story inspired). Take, for example, the debate that "This Guy Travelled The Country In a Pink Tutu Just to Make His Wife Laugh During Chemo" provokes. Commenters consider whether the manwho posts photos of himself in the tutudoes so really to amuse his sick wife, or more properly for the public attention. 38 Such responses to the uplifting anecdote, among others that it enables or inspires, will be discussed in this essay's final section. 39

Trolls and Tragedies
One way to track content's effects on the broader culturebeyond, say, the sociological surveyis to observe how its forms and ideologies infiltrate other media. Perhaps even more so than prior new forms of popular entertainment like TV, radio, or film, content is in a unique position to influence other types of cultural artifacts. Today's novelists, filmmakers, and TV writers, among others, do not simply consume content. They also create content and publicize their work through the content machine. Naturally, they translate content's aesthetics and concepts into their other work. Indeed, many are creating novels, films, or TV shows that emulate or respond to the uplifting anecdote. Katja Blichfeld and Ben Sinclair, for example, creators of the HBO show High Maintenance (which originated on YouTube), regularly write episodes that are extended uplifting anecdotes. 40 Literary author Teju Cole, after writing an essay critiquing an iteration of the content genre ("White Savior Complex"), has published a collection of captioned photography, Blind Spot, that (however inadvertently and distinctly) elevates the Humans of series, that original cyber compendium of uplifting anecdotes. 41 Here, I'll briefly discuss two works that, in addition to channeling the uplifting anecdote, illuminate some implications of its ethics: NBC's hit sitcom The Good Place and George Saunders's Booker Prize-winning novel Lincoln in the Bardo.
To begin with, the TV show and novel both take place in afterlife spacesthe "Good Place" and "Bardo"that are loose (and perhaps even inadvertent) allegories for social media. Those spaces, like social media, are virtual worlds in which persons of uncertain ontological statusneither dead nor aliveare drawn together through time and space. Here, as in Dante's Inferno or on Facebook or Twitter, questions of ethics are paramount: each character subjects his or her past life to moral scrutiny. Unlike Dante's ethereal beings, however, and more like today's Facebook or Twitter users, the denizens of the Good Place and Bardo are subject to not already decided, but rather ongoing ethical assessments. In the Good Place, residents are daily awarded a numerical score for the quality of their actions; as the series progresses, the significance of their old lives fades and their postdeath conduct takes on increasing importance. In the Bardo, the dead must embrace more selfless states of mind in order to ascend to some higher, unknown realm. Together, then, the TV show and novel allegorically symptomatize the same phenomenon: the fact that, today, social media platforms perform a quasitheological function. Surveilling eyes of digital "friends" motivate moral activity in much the same way that divine judgment once did.
More particularly, the TV show and novel, through their adaptations of the uplifting anecdote, reflect upon social media's distinct ethics. Each, to begin with, was in some literal sense inspired by an uplifting anecdote. Michael Schur, creator and director of The Good Place, attributed the idea for the show to a viral piece of content that he inspired. After a man filed an unnecessary insurance claim of $836 against Schur, Schur told the man that, if he dropped the claim, Schur would donate the money to Katrina victims. The story went viral and then morphed into a crowdfunding campaign (perhaps the only type of uplifting anecdote that demands some variety of real and not merely symbolic response). 42 Saunders, meanwhile, avowedly based Lincoln in the Bardo on a nineteenth-century iteration of the genre, a news story describing the tender manner in which Lincoln mourned his son Willie's death. 43 45 After he joined Facebook in 2011, he began to write fiction that resembled the uplifting anecdotes that regularly circulated on his page (e.g. an article relaying the story of the parents of the fallen immigrant soldier Humayan Khan, who were notoriously assailed by then-candidate Donald Trump). Many of his subsequent short stories have plots that resemble those typical of the content genrethe first and last stories of Tenth of December, for example, concern boys who save their neighbors from rape or suicide.
The Good Place and Lincoln in the Bardo also both incorporate the uplifting anecdote in other ways. Characters on The Good Place make frequent reference to such viral ephemera, comparing the ethically satisfying existence that the show's version of heaven promises to the feeling that such content evokes. Shortly after our protagonist Eleanor Shelstrop (played by Kristen Bell) arrives at the Good Place, she and the other new residents watch a PowerPoint style presentation delivered by their neighborhood's "architect" Michael (played by Ted Danson). The first version of this presentation is sponsored by "otters holding hands"; a second version compares living in the Good Place to "the way you feel when you see a chimpanzee and a baby tiger become friends." 46 Michael soon claims that his favorite flavor of frozen yogurt available in the Good Place is the comically abstract "when a soldier comes home to his dog." 47 Other examples abound. To name but one more: after Eleanor and her Good Place roommate, an ethics professor named Chidi, fall in love, they watch a video compilation of brief moments from throughout the show when one of them was particularly kind or affectionate with the other, à la Facebook's "friendiversary" videos. 48 In Saunders's Bardo, meanwhile, characters communicate with one another largely via short, status-update style monologues, many of which tell saccharine personal tales that resemble the uplifting anecdote. Indeed, many invoke the popular social media thematics of erotic or aesthetic recognition. On the novel's first page, for example, one of its two ghostly protagonists, Hans Vollman, tells the story of how he, a gentle old man, was wed to a young and beautiful girl.
Meek as a lambin sentimental and perhaps anachronistically "woke" fashionhe found himself unable to impose himself upon her sexually: "I could see her fear and . . . distaste." 49 Thanks to his kindness, she did eventually fall in love with him. But on the day when they are about to consummate their marriage, he became "sick" (or, as we understand, died). Next, Roger Bevins III, the novel's second protagonist, tells the story of how when he was a young man on earth, he finally surrendered to his homosexual impulses and took up with a lover named Gilbert. When Gilbert left, he attempted suicide. But half way through the act, in a classic sentimental trope far predating the internet (think It's a Wonderful Life), he realized how beautiful life was. 50 Other characters have similar tales to tell: Captain William Prince, after betraying his wife, repents and reforms; the young Elise Traynor, who died in adolescence, replays happy scenes with suitors, one of whom she might have married one day.
The novel and the TV show, moreover, depict worlds pervaded by the variety of ethics that social media in general, and the uplifting anecdote in particular, promote. Ethics, in the Good Place and Bardo, is ubiquitous and self-conscious but also anemic. In neither world are questions of ethics, to any meaningful degree, questions about real consequences; no action, up to and including murder, seems to be final. Indeed, in both worlds, characters can take almost no meaningful actions.
In the virtual world of the Good Place, everything done can be undone. In one episode, for example, Michael kicks a resident's dog into the sun; no matter, the beloved pet can be reconstituted from the ether. 51 Janet, the "helper," heartrendingly pleads for her life every time she is rebooted, but reappears, better than ever. The Good Place's four principal inhabitants -Eleanor, Chidi, Tahani, and Jason -end up, in the show's second season, living their first few weeks in the Good Place over and over, their memories wiped clean with each reboot. The merely virtual status of action, here, in addition to reflecting the show's allegorical relation to social media, underscores one of its broader theses: that people's power to do good, in a complex and contemporary world, is radically restricted. In the show's later seasons, the characters learn that earthly denizens' "goodness" scores are today lower than ever before, since almost no action can be taken that is not in some way enmeshed in a corrupt system. 52 Meanwhile, Saunders's ethereal protagonists are similarly impotent. They yearn to affect the earthly realm, attempting, by multiple means, to help Lincoln accept his son Willie's death (and thereby prevent the boy from becoming imprisoned in the Bardo). Ultimately, though, they must wait for Lincoln to find peace and acceptance on his own. Here, perhaps, in these characters' inability to influence the commander-in-chief is another reflection of the same contemporary sense of civic impotence.
Where real, consequential action is unavailable, ethicsas on social mediabecomes a matter of personal identity or self-expression. All that matters are inner states and the authenticity with which we outwardly represent them. On the first season of The Good Place, for example, the question of Eleanor's goodness quickly becomes the question of whether she truly means the positive actions that she takes. If, for example, she throws a party for her neighbors simply because she wants to earn goodness points, then the action is valueless: she earns none. If, however, she performs the action out of real altruistic feeling, then it will increase her score. 53 This is deontology by necessity; in a content-based world where all ethical actions are symbolic, they can have no value beyond the intent behind them. Indeed, the show's concern with intent is a direct product of its origins as an uplifting anecdote. Receivingeven solicitingso much digital attention for his own act of altruism made Michael Schur reconsider its ethical status. A New York Times Magazine piece describes his ethical progress via free indirect discourse: "He suspected that his mission was not, perhaps, entirely righteous. There was an element of grandstanding to the gesture, of moral one-upmanship, and Schur spoke about it with his family and colleagues and even professors of ethics. He became fascinated by the ways people can rack up ethical credits and debits." 54 For the denizens of the Bardo, too, "virtue" or the degree to which one is capable of earning a place in some rumored next world (presumably some sort of heaven, though we never learn) becomes a matter of the quality of one's inner life. The Bardo inhabitant's cardinal sin is not the one we would expect to see in some afterlife antechamber, or limbothe middling sum of one's good and bad ethical actions in life. Rather, their sin, again, is inauthenticity, a habit of dissembling born of inadequate self-knowledge. One reason that characters like Bevins and Vollman remain in the Bardo is that they cannot even admit to themselves and others that they are dead. Vollman refers to himself as still "sick" or in his "sick box"; whenever Bevins is on the verge of saying something like the word "dead," he omits it or is interrupted. 55 Meanwhile, Lincoln's son Willie is in danger of never ascending to the next world precisely because Lincoln cannot admit to himself that the boy is really gone. These characters will earn their place in the next world (or, in Lincoln's case, earn Willie's place), not by doing good deeds, but rather by achieving authentic self-expression: by knowing what they are and saying it plain.
In both worlds, as on social media, the equation of ethics with self-expression is also an equation of ethics with aesthetics or erotics. One of The Good Place's typical running jokes involves suggesting that major moral crimes and aesthetic faux pas are equivalent. The original list of actions that send people to the bad place, for example, commingles "genocide" (bad) with "use 'Facebook' as a verb," "ruin opera with boorish behavior," or "tell a woman to smile" (all also bad). 56 Other crimes cited throughout the show include liking the Red Hot Chili Peppers and "wearing jeggings as pants." 57 While these are clearly jokes, the show sincerely inhabits its aestheticized value system. On the level of its thematics, it never reaches a clear, philosophical verdict regarding what types of actions are good (beyond loving one's friends); on the level of its production, however, it clearly endorses an ethics of egalitarian representation. The diverse cast of four main characters, of different races and classes, are seamlessly drawn together in what YouTuber T1J has aptly called one of Michael Schur's "postracial dreamworlds." 58 The show also makes a point of ostentatiously working against televisual stereotypes, reminding us frequently, for example, that its Asian male protagonist, Jason, is both conventionally attractive and spectacularly unintelligent. The show also endorses a Platonic intermingling of the ethical and erotic realms, suggesting that the primary path to virtue is by loving and being loved (Eleanor becomes good by loving Chidi). Lincoln in the Bardo, as already suggested, also embraces the uplifting anecdote's elision of positive action and erotic recognition. As their personal, uplifting anecdotes suggest, the characters crave kindnesses from others that are amorous in nature.
Something is amiss, however, in these ethical worlds. Both The Good Place and Lincoln in the Bardo register the superficiality at the heart of their social medialike afterlives. Both the show and novel reach their climaxes when the protagonists realize that the worlds they occupy are not what they seem. In the final episode of the first season of The Good Place, Eleanor realizes that the Good Place is in fact the Bad Place. Though she, Chidi, Tahani, and Jason have been assuming that they inhabit a heaven designed for good people (with only Eleanor and Jason realizing that they do not belong there), Eleanor now realizes that the ethereal space is, in fact, actually an elaborate torture chamber of Michael's devising. Meanwhile, the denizens of the Bardo learn that, though they had believed that they were in some sense still alive, they are, in fact, dead. In both casesbut particularly in the case of The Good Place, which aired shortly after the 2016 electionthe collective anagnorisis seems to represent the recent revelation, on the part of educated liberal netizens, that social media was not what it had seemed. Far from a "Good Place," where liberals could gather together to share New York Times articles and display

C O N T E N T -E R A E T H I C S
316 or debate -their "wokeness," it was in fact a "bad place" where Russians and neo-Nazis spread malicious misinformation.
More particularly, these twin climatic turns imply dissatisfaction with the narcissism at the heart of social media's ethics. Eleanor's revelation punctures the assumption, on the part of the Good Place's denizens, that they are, in fact, so good. Tahani the heiress philanthropist and Chidi the ethics professorunlike the less idealistic, but more self-aware Eleanor and Jasonhave happily accepted that they belong among the heroic few. What they haven't realized is that their own goodness, like social media's prosumption-driven notion of virtue, has been as empty and insubstantial as the frozen yogurt on which they have been subsisting. Indeed, the fact that this airy confection was so abundant in the Good Place, the show jokingly implies, should have provided a clear indication of its nonparadisiacal nature.
Similarly, Saunders implies that the inhabitants of the Bardo, perpetually sharing their personal tales, have been mired in a pathological self-concern. The characters, more than anything, want to be rendered visible to those around them. "What did we want?" asks Vollman when Lincoln comes to see his son Willie for the first time. "We wanted the lad to, see us," he answers. To be seen, in this case, would be to live vicariously through Willie's having received an amorous or loving attention: "To be touched so lovingly, so fondly," says Bevins. "We were perhaps not so unlovable as we had come to believe." It would also, by extension, be to return from the dead. It's "as if one were -" Bevins says, omitting the implied word: "alive." 59 Understandable desires, to be sure. But for Saunders they turn out to be deforming. Each character is misshapen in a manner that corresponds to the story that he or she repeats, and its expressed desire for romantic recognition. Vollman, for example, has a gargantuan, floating erection; Bevins has many scattered eyes, ears, noses, and fingers, which represent his final, sensuous desire to evade his death. More to the point, obsessive self-expression, Saunders implies, is what keeps these characters stymied. "To stay," Vollman and Bevins together note, "one must deeply and continuously dwell upon one's primary reason for staying . . . one must be constantly looking for opportunities to tell one's story." 60 The Good Place and Lincoln in the Bardo take two different critical stances that social media's ethics compel, one of which -the novel'sis more promising than the other. The Good Place, at first, may seem to endorse proactive alternatives to social media's collapsing of the ethical and expressive. It suggests that though doing good is increasingly difficult, people might still manage to "try," even in the face of damnation ("we can try," Eleanor rallies her companions). Trying, by and large, will consist in forging companionate bonds with other persons, or honoring in a phrase that the show borrows from T.M. Scanlon -"What We Owe to Each Other." In reality, however, the position that the show most energetically embraces is not that of the sentimental contractualistthe version of Eleanor who exhorts her friends, fueled by love, to make an effort. Rather, it is that of the internet "troll"the version of Eleanor who punctures other characters sanctimony, rejecting their shallow, self-serving ethics in the name of a gleeful nihilism. Eleanor, by the show's estimation, begins a "bad" person. And yet her actions never truly offend. Her allegedly evil deeds, most of which involve puncturing the pomposity of self-important moralizers, are played for cathartic laughs. In one flashback, she refuses her boyfriend's request that they boycott a "problematic" coffeeshop; the meek man pales and wilts in the presence of her charismatic gusto ("I know you're going to break up with me," she says. "I saw the emails you wrote to your dad.") 61 In another scene, she refuses to participate in the nauseating group activities that her coworkers organize. 62 In a third scene, a heroically single-minded teenage Eleanor denies affiliation with every social group, including the so-called disaffectedshe immediately rebuffs the selffashioning revolutionary who approaches her. "It was so cool the way you told off those posers," he says. She responds with one word: "No." 63 Meanwhile, the only one of the four "bad" Good Place protagonists whose behavior seems actually designed to offend is Tahani. Tahani, despite having objectively accomplished far more good than Eleanor has ever done by raising millions of dollars for charity, is more palpably objectionable because of how smug she seems about it. When Tahani first comes to Eleanor's house with an elaborate, expensive housewarming gift, Eleanor displays her characteristic cynicism by mistrusting the gesture. And yet the show makes clear whose side we should be on. Tahani, the Instagram philanthropist, is the show's minor villain (at least in early seasons) because she believes that she belongs in the Good Place. Eleanor, the savvy spotter of cons ("I know liars," she says), is the show's social media-era hero because she knows that they both do not. 64 But where The Good Place, like Milton's Paradise Lost, is in spite of itself of the devil's -or of the troll's -party, Saunders's Lincoln in the Bardo is more earnest in its quest for an adequate, alternative ethics. To be sure, the book operates from within the world of social media's insubstantial morality, equating all virtue with self-discovery and self-expression. But it also often resists that prosumptiondriven ethics. It does so, moreover, without devolving into trollish nihilism. Rather, it attempts to re-focus public attention away from the heroism that inspires self-congratulation, and back onto the suffering that inspires action. In the moment when Lincoln finally accepts his son's death, the denizens of the Bardo, too, can see who they are: namely, that they are dead. Suddenly, and by implication as a result, they are released from the deforming prisons of their self-concern -Vollman and Bevins, for example, suddenly appear as normal human beings, without gigantic floating erections or thousand-fold eyes. 65 At the same time, Saunders revises the uplifting anecdote to similar effect. The novel, as we have seen, opens with Vollman and Bevins's sentimental tales. But these tales, as yet, have no endings. Vollman remainsat least in his viewon the brink of returning to the world to consummate his marriage. And Bevins, too, believes that he might return, perhaps to rekindle his affair with Gilbert. Saunders might have given these stories, like their social media counterparts, uplifting conclusions, granting these two characters the erotic recognition they desire. Instead, he does the opposite. Now, we suddenly learn a fact about which Vollman has been in denial: that his young wife has long since married another man, the "great love of [her] life," and she has already died and passed through the Bardo, thanking Vollman for putting her "on the path to love." Gilbert, we meanwhile learn, left Bevins for another man. At Bevins's expense, the two lovers "[shared] a laugh." 66 Here, the reader can feel no relieved self-satisfaction, as the scene inspires a far less complacent feeling: sympathetic pain at unredeemed loss.
New mass media, twentieth-century critics often complained, threatened the individual -Horkheimer and Adorno bemoaned the culture industry's homogenizing effects; David Foster Wallace critiqued television's capacity, among other things, to make the many one (he titled his essay on the topic, "E. Unibus Pluram"). In the social media era, a distinct sort of danger emerges: content, rather than threatening to overwhelm the individual viewer, insists on the perpetual presence of her symbolic self. As a result, it introduces a distinct sort of ethical terrain: a world in which all action is symbolic and expressive, and the measure of its quality is its perceived authenticity. From within this world, at least two types of critical stances, beyond the rejection of ethics (and sentiment) wholesale in the name of a more pressing politics, are available. There is that of the troll, which forecloses the possibility of ethical action still further; or there is that of the tragedian, which strives to open it up.
subscribers discover how many times pieces of content were shared on four platforms: Facebook, Twitter, Reddit, and Pinterest, during some specified period within the past five years. Users can use the tool to search by content domains (NYTimes), by article title, by author, or by title keywords (No blank search, returning the most shared content writ large, is permitted; this is also true on Facebook's similar CrowdTangle tool; that is, going directly to the Facebook API/Crowdtangle tool would not circumvent this issue). Restricting my searches to the period of July 2014 to July 2019, I searched, first, for content published by all of the domains listed by Newswhip as ever having produced the most shared content during a month between July 2014 to July 2019, and added all resultant pieces of content earning 500,000 or more total shares to my database. I then searched the content by generic title keywords (the top 100 most frequently used words in the English language, according to the OED, except "a" and "I" as one letter searches were not permitted), to get viral content from a broader sample of outlets; again, I added all pieces of content earning 500,000 or more shares to my database.
10 While focusing not simply on the headline text, but also on the text of the content itself would be, in theory, ideal, the task is unmanageable. Computationally scraping the individual pieces of content from the thousands of different content sources in my archive would be a task that would hardly take less time than doing so by hand. While it would be possible to scrape the viral content from only a handful of sources (e.g., only the NYtimes or Buzzfeed) in an efficient programmatic fashion, this would change the nature of the corpus in undesirable ways. It would turn the study from a study of viral content writ large into a study of viral content as produced by particular outlets.
11 "Incredible Dog Who Saved the Life of Newborn Baby Has Lasting Legacy," Resharewworthy, accessed October different numbers of topics (from 50 to 500, by intervals of 50), at multiple different optimization intervals (20-60, by tens). I used the standard stop-word list. I found that most of these experiments produced very good results (part of why I didn't need to adjust stop words). The topics were remarkably straightforward and easy to assign labels to, likely due to the simplicity of the documents. I decided that the most coherent list of topics were 150 topics at an optimization interval of 20. They are the topics on which this discussion is based. There are problems with the method of Topic Modelling, as many have discussed. For one thing, the model will produce slightly different results each time it is run. The topics produced should not be taken as an objective and comprehensive snapshot of all of the topics appearing in the corpus, but rather as a strong indication of some major themes and a starting point, on that basis, for closer analysis. Here, the topic model begins to indicate, to me, that the content divides into two major sorts of categories. Other methods like word counts and, not least, close looking, then confirm this suggestion.