The impulse to separate the literary critic from the mere consumer, the serious reader from the lightweight, is abiding. Several means of demarcation have been used; one is how much interest, and what kind, a reader shows in the person behind the work. The Guardian recently drew a line against the persistent interest in Elena Ferrante’s identity: Ferrante’s “true” fans, which is to say, the serious readers who grasp her work in all its depth, do not care about her identity. In the 1960s, a different literary trend insisted that the author was irrelevant. Ferrante refreshed that orientation, insisting that any attempt to learn about the author responsible for her novels was motivated by a kind of literary malfeasance, hardly higher than tabloid gossip. An author’s biography, Ferrante maintains, opens no window onto her work.
Perhaps. But history testifies that searching for the author responsible for a text is among our oldest impulses. And there’s a reason for that. In erasing the author, these theories have shortchanged us. The speed at which AI is learning to write like us should make us reconsider the human figure we’ve been so eager to dismiss.
In Toni Morrison’s Song of Solomon, the character Pilate Dead has no navel. Three years into her stay with a group of migrant workers, she lies down with a boy in the community, who then spreads a story of a stomach as smooth as stone. The community, once capacious and inviting, closes rank. She wasn’t “born natural,” they say, and so they kick her out. Pilate learns to turn her strangeness into a source of power. Morrison said of her character: “The navel allowed me to introduce the fantastic early in the book. It meant that Pilate could be ‘inside and outside’ at the same time,” adding, “She was ‘innocent wisdom.’” With no umbilical scar connecting Pilate through an unbroken line back to Eve, she carries the wisdom of the forbidden fruit without its mark of shame.
But as we saw in the migrant workers’ reaction, the same condition can signal unnatural wisdom, the uncanny, even the monstrous. In the TV show Murderbot, Alexander Skarsgård’s character, a cyborg security agent called a SecUnit, also has a smooth stomach: a mark of his syntheticism. A row of unclothed SecUnits in storage is like a row of apples with no indentations where a stem once was. Perfect, dreadful smoothness all around.
It is foretold that AI will produce works of literature of such quality that we’ll be unable to tell the difference between it and the work of human hands. These works will lack navels in practice, but, since AI models were created by us in our own image, they will have navels of a synthetic sort, suggesting a trail of origin that is not detectable, that has dematerialized into a folder on a server.
Should we be concerned? Does having a navel — the old-fashioned kind, made of flesh — matter?
Ferrante’s best-selling Neapolitan Quartet, which serialized the upheavals of midcentury Italy through the lives of its protagonists, was covered as much for its author’s refusal to be known as for its literary merits. She famously summed up her belief in a private letter (that she then published): “I believe that books, once they are written, have no need of their authors.” Ferrante practiced what she preached — sort of. From the start, people wanted to know about the author behind the books, and from the start, she held up one hand as a barrier and with the other beckoned them forward, dangling scattered self-disclosures that drew curiosity.
This led to an inevitable climax: in 2016, the Italian journalist Claudio Gatti concluded after an investigation that Ferrante was the literary translator Anita Raja. When the New York Review of Books broke this news to English-speaking audiences, others in the literary establishment hurried to distance themselves from the announcement. “For Literary World, Unmasking Elena Ferrante Is Not A Scoop. It’s A Disgrace,” ran an NPR headline. Even the New York Review of Books, after receiving “many queries and comments,” pinned an apology to its original coverage.
The New Yorker’s coverage in particular oozed distaste. “I hate to do it,” wrote Alexandra Schwartz, before recapping Gatti’s findings. She explained how Gatti tracked down income payments, tax records, and property purchases, to puzzle together that Raja, who supposedly earned income only from translating, must have had another substantial source of funds. She recast Gatti’s fairly straightforward sleuthing as classist, attributing to Gatti the view that Raja had “risen suspiciously above her station.”
The controversy sprang from a tangle of textual purity, privacy, and feminism. From the Gatti fallout, it was clear that the media was expected to honor a privacy pact with Ferrante despite never having agreed to one. Schwartz, like “most” Ferrante readers, in her estimation, had relished meeting Ferrante on “an imaginative neutral ground,” where she didn’t know anything of Ferrante and Ferrante didn’t know anything of her. Gatti had destroyed that protected space.
But did it ever really exist? Despite her protests that authors were harmful distractions, Ferrante did not stay silent about herself. Schwartz wrote that Ferrante “has chosen to be anonymous,” but that wasn’t true. Ferrante assumed the name of an Italian woman and used that name across a body of work, deciding that some things, at least, were important for readers to know. As her renown grew, Ferrante went further still, periodically dripping biographical snippets and personal opinions across interviews with such publications as the Paris Review and in columns and essays for the Guardian and the New York Times. Many of these self-disclosures were later collected in Frantumaglia, a plump volume of Ferrante’s interviews and private letters which she and her longtime publishers assembled. Given the furor around the unmasking — fanned by Ferrante’s publishers on her behalf — it’s hard not to see Frantumaglia as a cynical exercise in telling readers how, and how not, to read her works, while disclaiming any such motive. The heart of the matter for Ferrante was not anonymity but control: the right to dictate the terms on which she would be known.
Her pseudonymity was a sword, not a shield. Rather than snuff out her authorial “aura,” as she claimed to want, Ferrante produced something more powerful: a tantalizing veil, a mystique that inspired devotion. Readers did see an author behind the work. What differed in this case was whether they pledged allegiance to the vision Ferrante presented or peered more deeply behind it.
Whether you agreed to her terms or kept digging for clues was not a neutral choice, and it was Ferrante herself who set the stakes. In an interview, she mused, “Who really cares about the person who wrote it?” (She’s given dozens of written interviews.) Elsewhere, she brushed off this curiosity as the concern of “only the narrow circle of those who work in the media,” people who “either are not readers or are hasty readers.” “Like it or not,” she told the interviewer, “you, because of your job and apart from your sensibility as a cultured person, feel called on to fill in the space that I purposely left empty with a face, while readers fill it in by reading.”
That response typified Ferrante’s tenor when such questions arose: condescending dismissal. The fevered reaction even to secondhand accounts of Gatti’s investigation, like NYRB’s reporting, can be partly traced to the divisions Ferrante had set up: between those who are cultured and those who aren’t; between those who are readers and those who don’t merit the label. In her taxonomy, only the uncultured, unreaderly group was curious about Ferrante. This frame was uncritically taken up by people whose profession is to criticize. Perhaps afraid of being thought unserious, those who had built reputations on their judgments of literature — on their status as cultured readers — fell in behind the line Ferrante had drawn.
Of course, Ferrante has the right to play with pseudonymity and self-presentation. But the collective pretense that such a game was not being played is not justifiable.
Ferrante was not the first to play this game. She was not breaking new ground in relegating the author, even if she carried the idea to a wider audience. Her view mirrors a midcentury movement — born when her Neapolitan protagonists were coming of age — that reshaped literary criticism.
“The author is a modern character,” Roland Barthes declared in his 1967 manifesto “The Death of the Author.” Literature, he argued, had been trapped in the tyranny of the author for too long. “Explanation of the work is still sought in the person of its producer, as if,” Barthes wrote, “it was always, ultimately, the voice of one and the same person, the author, which was transmitting his ‘confidences.’” “To assign an Author to a text is to impose a brake on it,” and so the very existence of authors is destructive. In the shadow of its author, a text lies frozen, inert — goes the theory.
Like manifestos generally, this one rested on more than a little sleight of hand. Barthes’ argument purported to flow from history: in pre-modern, shamanic, “ethnographic societies,” he claimed, there was no concept of a singular author. But he gave no specifics, offering only the jargon of anthropology without the burden of evidence. Still, his claim that the author was a modern invention became a discursive landmark that later theorists had to reckon with.
Two years later, Michel Foucault, in his essay “What Is an Author?” posed a related question: when did the author-function, the one that organizes, classifies, and confers authority on a text, emerge? Barthes’ aim was to transform criticism and his mode was polemic — Foucault’s object was more measured: to understand the functions of the author so as to ask better critical questions. Still, like Barthes, Foucault traced the emergence of particular aspects of the author-function to post-Enlightenment Europe, when modern systems of intellectual ownership and property rights were codified. If a specific cultural era created our author, then contemporary culture could un-create her.
There’s a line from Beckett on which Foucault anchors his essay: “What matter who’s speaking, someone said, what matter who’s speaking.” The line summed up the emerging critical position that the significance of a text lay only in its words, not their origin. That was the belief that Barthes’ manifesto sought to enforce, and one of his enforcement tactics was to link the author to “capitalist ideology,” to “empire.” The implication for his 1960s audience was that to appreciate the author was to out oneself as a reactionary.
To Barthes’ idea that the author is a modern invention, a different Beckett line, from the novel Murphy, supplies a corrective: “The sun shone, having no alternative, on the nothing new.” This line operates both as an assertion and as proof of itself by reworking the famous line of Ecclesiastes: “There is nothing new under the sun.” The same sun beams down on the author, who is “nothing new,” but ancient.
In classical Rome, the author was already alive and well. Suetonius wrote biographies of the poets. Critical commentaries read the Aeneid against the history of its author, Virgil. Authors weren’t only figures of prestige; they could be notorious, too. Augustus banished Ovid to a remote town on the very edge of Roman civilization for carmen et error: a poem and a mistake. Before modern IP law, there existed a moral intuition that authors had distinct claims to their creative work. In 86 CE, Martial attacked Fidentinus for passing off Martial’s poems as his own. Martial called him a plagiarius — a kidnapper — the etymological forefather of “plagiarism.” The concept predated Fidentinus’ deceit. The ancient Greeks hurled charges of literary klopē — theft — when one writer was suspected of using another’s words without attribution.
Further in the past, reverence for authors thrived in ancient China. In the fifth century BCE, Confucius was widely revered as an author; his name bestowed prestige on the writings (sometimes falsely) attributed to him. In China’s classical Warring States period, Qu Yuan composed the first-person poem Li sao (“Encountering sorrow”) in a subjective, hallucinatory style. His own story was tragic: once a courtier, he fell victim to jealous slanderers, lost the king’s favor, and committed suicide. Qu Yuan’s contemporaries read his poetry with his dramatic past in mind. As Mark Edward Lewis explains, in Writing and Authority in Early China, “text was bound to the narrative of a presumptive author’s life and understood as a record of his experiences, so no reading of the poem could escape reference to the poet.” It was a Barthesian nightmare.
Skipping east across the sea, the Heian tradition of Japanese poetry took for granted that works had individual authors. In the late 800s, Ono no Komachi composed this poem:
Although my feet
Never cease in running
On the path of dreams,
Such nights of love are never worth
One glimpse of you in your reality.
This is not an epic, authorless style. And we presume to our discredit that the poem’s audience, laden with knowledge of the poet’s identity, felt constricted to a singular meaning. Poetry during this time was memorized and recited: it would have been customary for a lover to use Komachi’s words — or another poem from their library of memory — to speak their own passion. Komachi did not have to die for her poem to live.
To be sure, the role of authors has not been perfectly fixed across cultures and centuries. The particularities have varied, as have the metaphysics: Barthes and Foucault are part of a line of scholars and philosophers seeking to understand who, exactly, this author person is.
But the notion that the author’s emergence can be calendared is fanciful. There are certain things that arrive not as pinpoint discoveries, like gravity or perspective drawing, but have been encoded in the soupy mix of our genes, like an aversion to spiders and a rough-and-ready notion of fairness, waiting to unfurl when conditions are right. A more considered view places the author within this mix.
What Barthes left out of his account, and Foucault gestured toward but didn’t fully arrive at, is that while the work is the creation of the author, the author is the creation of the reader. The author is a mediated entity — not simply the person who composed the work, but someone arising through the work, refracted through biography, rumor, reception, resistance, devotion.
Because the trail doesn’t stop at the classical traditions of Rome, China, and Japan. As long as there have been readers — or earlier still, hearers — of another’s words, there have been authors.
In Genesis, God spoke the world into existence. This verbal creation motif is widespread: in the Rigveda and later Hindu philosophy, in ancient Egyptian theology and more, creation is tied to utterance. Across traditions, words are the substrate of reality, and the figure of the author stands as a metaphor for divinity. If you step back far enough and squint, you’ll glimpse a pattern: we read the world and inferred an author, and that author is God. The author-concept is so fundamental to human cognition that we projected it onto the cosmos itself.
Science has long reckoned with why and how the universe came to be. Why is there something and not nothing? The point at which the first atoms emerged is an occasion of wonder. But big bangs are happening all around us, continually. Each work of literature — each Song of Solomon, Pride and Prejudice, Tale of Genji, Through the Looking-Glass — used to be nothing, and then: something. The law of conservation says that matter is neither created nor destroyed, only changed from one form to another. Yet a work is not made of matter and is ungoverned by this law. Pemberley weighs nothing. Alice takes up no space — her Wonderland takes even less. Before she was Pilate Dead, she wasn’t something else. She just — wasn’t.
Where this kind of creation is concerned, science so far offers little explanation; the human mind remains a place of great mystery. A brain is gray and lumpy — three pounds of frankly unappealing matter you could hold in two hands. Yet from this gray lump springs vistas that are boundless, saturated, urgent, mystical. Through reading, we are transported even as we stay in one place, and the people we encounter on these journeys — the inventions we call characters — may seem more real than people of flesh and bone.
The death-of-the-author denies this wonder. A novel is a product of its time and of times that came before it, mediated through a unique gray lump. Moby Dick was not waiting in the ether for someone to pluck it down; it emerged from one person’s particular interior. And yes, a reader makes meaning — but from something that is there already. As readers, we weren’t around when nothing became something. Like the recorders of the Old Testament, we saw what came after. We saw the work, and through the work, we imagined a creator — the author.
Ferrante’s account considers irrelevant that the author of Song of Solomon, Sula, and The Bluest Eye was a black woman born in Ohio in 1931. That the author is a specific person, Toni Morrison, writing from a specific perspective. “Today, who really cares about the person who wrote it?” Ferrante has asked. Yet if a work is successful, then it serves as a prism through which truth is refracted. And the author’s position determines the angle at which light enters, and how it bends.
Text without context risks sterility. Zadie Smith suggests why: it is the presence of a consciousness — the author’s own — that gives writing its vitality. If no sense of an author bleeds through, Smith explains, “if the sentences are from a template offered up by the wider culture and reveal nothing of the consciousness behind it all, then that novel will mean very little to me.” She wrote those words in 2019. Today, “template offered up by the wider culture” reads as a prophetic description of large language models.
We do not read only to be entertained. We read to be moved, and the power to be moved lies in the work’s ability to tell us something true. We want a link to a real human being — or as Smith has it, a real human consciousness.
We have all read novels that enveloped us in states of wonder. And if they had been written by artificial intelligence? Taking the Barthes-Ferrante view to its logical extension, that’s irrelevant, for after all, it doesn’t matter who’s speaking. But it matters a great deal. What is said takes a different shine depending on who’s doing the saying. That’s a foundational premise of fiction itself, which is narrated: The voice, not just the words, matters. For that premise to hold any truth within the imagined world of the novel, it must have a real-world analogue. Fiction succeeds only if, by lying, it tells us something true about the world.
Consider Wuthering Heights, which was initially published in 1847 under the pseudonym Ellis Bell. When Emily Brontë died in 1848, Charlotte became her literary executor — and reintroduced the novel under her sister’s real name. What Charlotte wrote in her introduction is one of the most illuminating acts of author-to-author reading in the English literary tradition: a grieving sister who understood better than anyone how the woman and the work were inseparable.
In it, Charlotte granted certain criticisms of her sister’s novel, rooted in its “rusticity.” As a reader, Charlotte wrote, “I admit the charge, for I feel the quality.” In other words, the work was imperfect:
It is moorish, and wild, and knotty as a root of heath. Nor was it natural that it should be otherwise; the author being herself a native and nursling of the moors. . . . Had [Emily been] accustomed to what is called “the world,” her view . . . would have differed greatly from that actually taken by the home-bred country girl.
Imperfection is its own site of meaning. Charlotte grants the charge of rusticity, then flips it: the book is uncivilized and tangled because its maker was nursed by an unconquered wilderness. A navel-string connects them all: nature, nursling, novel. A more worldly author might have produced something “wider — more comprehensive,” but, Charlotte added, “whether it would have been more original or more truthful is not so certain.”
Readers don’t need permission to care about authors. But for those who demand justification, here it is: provenance as integrity. Emily’s wild sensibility is not additive but constitutive — the grain of the work itself.
The text is one thing; the human spirit that created it is another; both are precious. It doesn’t mar the experience of reading Wuthering Heights to discover something of Emily Brontë’s life. It enhances it. Maybe Ferrante and her adherents would say the story of human creation is beneath the focus of proper readers. I say Ferrante has it backwards.
Barthes argued that if we keep the author in mind, we cannot see a work on its own terms. This is why he wrote his manifesto: to kill the author so that the reader might live freely. Critical reception of Ferrante seemed to validate this view. It was through “Ferrante’s insistence on total privacy,” Meghan O’Rourke argued in the Guardian, that we focused on “her sentences, the patterns of her mind, the path of her imagination.” She continued: “there is so much continuity of thought from novel to novel,” and only “in her biographical absence are we permitted to focus on that literary continuity.” O’Rourke plainly implies that knowledge of an author destroys our ability to attend fully to a text.
There was one reader who loved Wuthering Heights and Jane Eyre. She made casual references to their authors in her letters, as if the Brontës lived down the lane. In a condolence letter, her way of honoring the departed was to compare them to the “gigantic Emily Brontë, of whom her Charlotte said ‘Full of ruth for others, on herself she had no mercy.’” She devoured written accounts of the sisters, declaring the 1883 biography of Emily Brontë as “more electric far than anything since Jane Eyre.” She wrote her own poem in tribute to Charlotte, and she asked that a poem by Emily be read at her funeral, a wish that was fulfilled. This reader, Emily Dickinson, cared about the authors behind the works she loved.
Yet if anyone ever knew that the poetic “I” is not the same as the hand that writes, that a text encodes multiple meanings and is not identical with its author, it was Dickinson. A century before Barthes, she explained her first-person poetry to a friend: “When I state myself, as the Representative of the Verse — it does not mean — me — but a supposed person.” Nor did Dickinson view her own meaning as definitive. Once a poem was set loose, she knew, it lived a life beyond her, as she put it in poem #1212:
A word is dead when it is said,
Some say.
I say it just
Begins to live that day.
The death-of-the-author doctrine rests on two notions: that authors would force single-minded interpretations of their texts, and that readers must be freed from such tyranny. Dickinson puts the lie to both. As an author, she knew that her work’s meaning went forth not fixed, but alive. As a reader, she loved the Brontës — her conception of them — with a fierceness that reflected the personal meaning she found in their work. Dickinson did not have to choose between caring about the author and respecting the work’s independent power, and neither do we. We can be curious about Dickinson herself, admire her idiosyncratic brilliance, look for clues to the verse in the life of its author, and still find our own meaning in her poetry.
We did not need Barthes or Ferrante to rescue us from our supposedly helpless dependence on the author. Rescue is needed, though, from the patronizing notion that readers can’t be trusted to approach the author as a companion rather than a dictator, and so the very notion of the author must be banished entirely, as if we are children and the author is refined sugar.
Barthes and Dickinson were spared the dawn of LLMs. AI will never be a native and nursling of English moors, nor a granddaughter of sharecroppers, nor the situated offspring of any other particular time, place, and people.
Its origins resemble a manufacturing process called sintering wherein a powder of countless particles is compacted and heated until the particles fuse at their edges into a single mass — giving the false appearance of wholeness. An AI text might look right on the surface, but it’s a simulacrum, smooth and serviceable but never quick with life. Work produced by AI has no grain: the navel-string running from the text back to a particular body in a particular place. A sintered text has a billion mothers, which is to say, none.