“Manuscripts don’t burn (рукописи не горят).” — Mikhail Bulgakov

The announcement of the July release of Go Set a Watchman, Harper Lee’s sequel to To Kill a Mockingbird, first met with delighted surprise. There had been no reason to believe the famously reclusive Lee would ever publish again; when asked if she was writing a follow-up to her tremendously successful novel, Lee is reported to have replied “I said what I had to say.” It seemed there would never be more of her work than Mockingbird and five nonfiction essays — until one day, when her lawyer Tonja Carter went to an archive to check on the original manuscript of Mockingbird and discovered a complete novel “believed to have been lost for more than 50 years.”

The response to the appearance of Watchman was a kind of joy classical scholars are familiar with from discoveries of long-lost poems by Sappho and Simonides. We know how it feels to long for more material by a beloved writer. The magnitude of that longing is such that when an unexpected addition to the classical corpus appears, we examine it with extreme suspicion. We know that whenever something is desired to that degree, it presents a temptation to lie, steal, or fabricate.

So it came as no surprise that the narrative surrounding the announcement of Watchman soon took a turn. Suspicions arose that Lee’s publisher, HarperCollins, was taking advantage of her to capitalize on the tremendous success of Mockingbird. Journalists pointed out that the timing of the discovery was suspicious, coming only months after the death of Alice Lee, Harper’s sister, lawyer, and longtime advocate. The ‘long-lost sequel’ became a draft Lee had never finished and never intended to publish. Rumors circulated that a stroke in 2007 had left her blind, deaf, and forgetful — unable, in short, to consent to Watchman’s publication, let alone participate in it.

If these suspicions are correct, Watchman resembles less the Verona Codex of Catullus found under a beer barrel than the Aeneid, published by Augustus even though Vergil allegedly asked his editor to burn it. There are a surprising number of other stories about great works of literature that would have become kindling had their authors’ wishes been respected, including most of the works of Emily Dickinson and Franz Kafka.

The stakes are high for Go Set a Watchman, perhaps nearly as high as they were for the Aeneid. Vergil was Augustus’ chosen poet, writing an epoch-defining instant classic; Harper Lee is the author of what many consider to be the American novel. The publishers of Watchman must know what Lee means to readers, and they have tried (perhaps not satisfactorily) to combat rumors that the novel is being published without her consent and participation. HarperCollins would probably like it very much if, when the book is released next week, the press would return to the ‘long-lost sequel’ narrative they initially constructed. But perhaps they should have embraced the uncertainty.

If I’m being honest with myself, I’m not sure I could have ever judged Watchman on its own merit. How could I read about Scout Finch, returning to her home older and wiser, without really feeling like I was reading about myself, returning as a more experienced and knowledgeable reader to a book I loved in middle school? Whether I responded with nostalgia or disappointment, that feeling would be about me as much as it was about Lee’s book — probably more. The rumors surrounding Watchman’s discovery only increase that potential: if the book is being read in a manner that Harper Lee didn’t intend, in a way it isn’t entirely hers. And if it isn’t hers, I have a greater freedom to feel it’s mine.

Readers have always had an undeniable fascination with texts that authors didn’t consider quite finished. We feel a voyeuristic thrill knowing we’re going against the author’s wishes by reading something they didn’t want to be read. The relationship between writer and reader is typically an orderly one, mediated by the text itself. But when the writer asks that a book be burned, its subsequent publication creates an unregulated and anarchic world of interpretive possibility.

If I told you I wasn’t ready to publish this essay, would you keep reading?

The most detailed version of the story about Vergil’s deathbed wishes comes from Suetonius, by way of Donatus:

“Before leaving Italy, Vergil had arranged with Varius to burn up the Aeneid if anything befell him; but Varius had insisted that he would not do so. For this reason, when his health was failing, Vergil demanded his scroll cases earnestly, intending to burn them himself… Nevertheless, Varius published them, on the authority of Augustus, but revised only in a cursory fashion (summatim emendata), with the result that he left unfinished verses, if there were any” (trans. Wilson-Okamura via O’Hara)

Most classicists disbelieve the ancient biographical tradition almost as a reflex. Although this skepticism is often warranted, we can feel near certainty about two elements of this story. The first is that the Aeneid is unfinished: it contains 58 lines that do not form a full dactylic hexameter. That only constitutes half a percent of the poem’s roughly 10,000 lines, but it still suggests Vergil died before he was able to complete the Aeneid to his satisfaction.

Second, although we can never know what Vergil’s dying wishes were, the story about him wanting the Aeneid burned is very old — so old that it may have begun circulating within years of his death in 17 BCE. In Tristia 1.7 (8 CE), Ovid claims his exile from Rome prevented him from polishing the Metamorphoses to the high gloss he would have liked (defuit et scriptis ultima lima meis, 1.7.30), so he threw them on a pyre (1.7.19–20). But the books survived nevertheless, and he asks that a preface be added excusing their roughness (1.7.35–40).

Tristia 1.7 is generally accepted to be part of Ovid’s project to assimilate his (death-like) exile to Vergil’s final voyage to Greece and the Metamorphoses to the Aeneid. (Nita Krevans’ account of the productive dissonance in this assimilation is excellent.) Ovid claims that nobody reading the Metamorphoses closely could fail to observe how he had been unable to put the finishing touches (summam… manum, 1.7.28) on it. This sentiment came as a surprise to me, since I’d always considered it one of the most highly polished pieces of literature I’d ever encountered.

In spite of these protestations, Ovidian scholars have treated the poem as complete, and Vergil scholars likewise have chosen for the most part to believe that the Aeneid was essentially finished. James O’Hara rejects analyses that attempt to understand “inconsistencies” in the text as evidence of incompleteness and argues that these inconsistencies — and the poem’s disconcertingly abrupt ending, which lacks any reference to the founding of Rome or the marriage of Aeneas and Lavinia — are part of Vergil’s poetic technique. He writes (p. 96):

“There are indeed signs that some passages may be lacking final stylistic revision, but often when readers claim that the poem is unfinished they reveal less about the poem’s flaws than they do about their own assumptions about life and literature, and the ways in which certain inherent qualities of the poem are incompatible with these assumptions.”

But couldn’t something similar could be said about all literary analysis? It’s certainly true of the scholars who go too far in the other direction and argue that Vergil’s half-lines are also part of his poetic technique. A few half-lines do seem more powerful for their brevity and the sense of incompleteness the metrical anomaly creates: Aeneas disavowing his free will in leaving Dido (4.361), Andromache trailing off while asking whether Ascanius still lives (3.339–40). But O’Hara argues that any added poignancy is an accidental benefit rather than an intentional innovation: “For every apparently effective half-line, there are dozens that add nothing to the poetic effect even when looked at from the most experimental modern perspective” (p. 100).

There is no evidence to believe other than that, had he lived longer, Vergil would have replaced the half-lines with complete hexameters. Those who argue otherwise, like those who argue that the Aeneid is rough around the edges, tell us more about themselves than they do about the poem. The latter group reveals that they would rather risk engaging in the tenuous, slightly insane imaginative process of trying to deduce what Vergil’s intended ending was than accept that the poem denies us the satisfaction of closure. The former group reveals an unhealthy dependence on authorial design. For these scholars, the unexpected beauty of some of the half-lines can only be valid if Vergil intended them to look as they do.

To anyone trained (as I was by one of my dissertation advisers) not to fall into the trap of the intentional fallacy, this entire discussion is fraught with dangers. Although the textual critic is an intentionalist, sorting through variant readings to recover the text as the author originally intended it to look, the literary critic is supposed to accept that authorial intent can never truly be accessed and is therefore more or less irrelevant. (Many authors find this idea repellent, which is perhaps why they are so prone to destroy their own works before they can reach any readers.) All that matters is what can be understood from the text itself (if you prefer new criticism) or what the text means to a given reader (if you prefer reader-response criticism).

One translator of the Aeneid, Howard Felperin, recognized the potential in the uncertainty of Vergil’s text:

“The air of indeterminacy that haunts an uncompleted poem, hovers about its margins as it were, invites a freedom in dealing with it that we wouldn’t feel in the presence of perfection. It might even inspire and license, though by no means underwrite, our own limited authority in the face of greatness” (pp. 20–1)

To treat the Aeneid as finished is to ignore the evidence that Vergil intended it to look (slightly) different than it does. But accepting that evidence creates a shadow-text — a complete, perfect, Platonic form of the Aeneid of which the received epic is a degenerate form. That this shadow-text is unknowable does not prevent us from imagining myriad possible shapes for it. To textual critics, it can be a relief to brush the shadow-Aeneid aside — to treat the text we have as so close to finished as to be nearly indistinguishable from the Aeneid-form — and bask in the warm, comforting glow of Vergil’s design. But literary critics sometimes find the shadow-text’s indeterminacy more enticing.

If Vergil truly did ask that the Aeneid be burned, he is part of a long genealogy of writers who have made similar pleas. Emily Dickinson’s will instructed her sister Lavinia (life imitates art) to destroy her papers, a request Lavinia interpreted loosely to include only her sister’s correspondence and not her poetry. Franz Kafka, thought to have burned as much as ninety percent of his work during his lifetime, wrote to the executor of his will, “Dearest Max, my last request: Everything I leave behind me . . . in the way of diaries, manuscripts, letters (my own and others’), sketches and so on, [are] to be burned unread.” Had Max Brod listened, The Trial and The Castle would never have been published. Vladimir Nabokov also requested that the unfinished manuscript of The Original of Laura comprising 138 index cards be burned.

When a writer asks that their manuscripts be incinerated, they are denying anybody else the authority to edit and publish their work. Scholars and commentators wilfully ignore that authorial desire every day, of course; but the unfinished text presents a peculiar kind of temptation. When the author asks that it be destroyed, they attempt to exert control over its (non)transmission — but instead cede the sole ability to shape its final appearance.

Or, to rephrase: if the story is true, then according to Vergil, the first and second rule of reading the Aeneid are both “do not read the Aeneid.” Every reader and scholar of the text is already breaking the rules.

Where Dickinson and Kafka paralleled the Vergilian deathbed narrative, Nabokov actively explored it. His death in 1977 prevented the completion of The Original of Laura, and although he requested that it be destroyed, his wife and son published it in 2009. Nabokov himself explored this scenario with astonishing prescience in Pale Fire, a novel (of sorts) that embodies the worst nightmare of every writer and commentator.

Pale Fire has four parts. The second is an unfinished 999-line poem (also called Pale Fire) in rhyming couplets by the murdered poet John Shade. The poem explores primarily Shade’s responses to both the suicide of his daughter Hazel and his own near-death after a heart attack. The other three parts of Pale Fire are the foreword, commentary, and index to Shade’s poem, all written by his neighbor Charles Kinbote. Although Kinbote writes ad loc. 47–8 “I have no desire to twist and batter an unambiguous apparatus criticus into the monstrous semblance of a novel,” his ‘commentary’ tells a different story. It is mostly his narration of the story of the daring escape of King Charles Xavier the Beloved from the “distant northern land” (p. 315) of Zembla. It swiftly becomes clear that although Kinbote means to gradually reveal to the reader his true identity as King Charles, instead he succeeds only in gradually revealing to the reader that he is completely insane.

But dismissing Kinbote as delusional is easy. It is less easy to admit that most scholars — if not all! — share in his delusion to some degree.

Kinbote rather immodestly suggests (p. 28) that the reader ought to read his commentary first, then read it alongside the poem itself after cutting the pages of the poem out for ease of back-and-forth reference (of course, today you can just bring up the text of the poem on your computer), and then finally read the commentary alone a third time. (I can hardly imagine any reader following these instructions, though they do remind me of recommendations by some of my graduate school instructors on how to read a commentary properly while preparing for exams.) He claims a tremendous authority over the poem (pp. 28–9):

“Let me state that without my notes Shade’s text simply has no human reality at all since the human reality of such a poem as his (being too skittish and reticent for an autobiographical work), with the omission of many pithy lines carelessly rejected by him, has to depend entirely on the reality of its author and his surroundings, attachments and so forth, a reality that only my notes can provide. To this statement my dear poet would probably not have subscribed, but, for better or worse, it is the commentator who has the last word.”

That last sentence should send a chill up every writer and scholar’s spine; it is the precise scenario Donatus’ Vergil seems to have been trying to prevent.

As a scholar, I often felt more amused than professionally offended by Kinbote’s deranged commentary. Lines 61–4 of Shade’s poem read:

TV’s huge paperclip now shines instead
Of the stiff vane so often visited
By the naïve, the gauzy mockingbird
Retelling all the programs she had heard

Kinbote’s commentary on these lines (p. 95) reads “Line 62: often — Often, almost nightly, throughout the spring of 1959, I had feared for my life….” The distance between poem and commentary is so immense that the two barely seem to belong between the same covers.

I did, however, feel offended by Kinbote’s obvious disdain for anything resembling real scholarship. He finds the task of commenting on and providing information about Shade’s dead daughter tedious; he would rather write about “richer and rarer matters” (p. 164). Even worse is his treatment of a pivotal moment in Canto III where Shade recounts how, during his heart attack, he had a vision of Old Faithful (l. 707, 716, 745). After seeing a magazine story about a woman (Mrs. Z.) who had a similar vision of a fountain during a heart attack, he attempts to contact her, hoping for confirmation that he had truly glimpsed the great beyond. Instead, he discovers that the article contained a typo: Mrs. Z. saw a white mountain, not a fountain. His attempt to find meaning is frustrated. It might be one of the best moments of the entire poem, and Kinbote’s commentary reads (p. 256), “Anybody having access to a good library could, no doubt, easily trace that story to its source and find the name of the lady; but such humdrum potterings are beneath true scholarship.”

Kinbote’s analysis isn’t precisely in keeping with the school of reader-response criticism. He actually means that the poem cannot be understood without background information only he can provide. Kinbote claims to have influenced and inspired Shade to create a poem on his Zemblan theme. That the poem contains no direct reference to story of King Charles Xavier is no deterrent: Kinbote can hear the echoes and “subliminal debts” to his tale (p. 297).

Scholars have argued variously that Kinbote murdered Shade himself (in spite of the careful construction in his commentary of the narrative of the approach of the would-be assassin Gradus); that Kinbote wrote the entire book, including the poem attributed to Shade; or that Shade wrote the entire book, including the material attributed to Kinbote. Any of these readings could be right, or none. Pale Fire represents an irresistible temptation to scholars because it speaks to the problematic nature of scholarship. As a book about the dangers of overly imaginative commentary, it both invites and rejects the possibility of overly imaginative readings of itself. It is difficult to read the book without becoming a second-order Kinbote.

But the process of analyzing this book should be less one of becoming than it is of realizing you have been Kinbote all along. What commentator doesn’t think their extensive knowledge of a work of literature — perhaps more than any other reader in the work’s history! — is what entitles them to write a commentary to guide the reading of others? Haven’t we all searched for hidden, subtextual meanings, as though the author is whispering a message we alone can hear? I have spent so much time immersed in a text that I begin to feel it is part of me — and, in an odd way, I am part of it. When I write about that text, I am really writing about myself. My experiences, interests, and history determine the texts I am drawn to and shape my interpretations of them.

As I write these words, I realize: this essay is, at some level, really about me. It is about Harper Lee’s sequel, Vergil’s deathbed wishes, Kinbote’s insane commentary, and what I think is at stake in being a scholar and a commentator and a writer. Everything I’ve written in the past year (or perhaps longer?) has been about trying to answer the same questions. Why do scholarship? Does reading and writing about other people’s work help me know literature, or myself? If it does both, then my literary analyses are in some sense autobiographical: does that subjectivity invalidate them? Am I, in the end, no better than Charles Kinbote?

One of the last lines of Pale Fire is Kinbote’s assertion, “John Shade, perhaps, will not be too much annoyed by my notes” (pp. 298–9). Little though we might share this belief, it is a revealing statement: even Kinbote, at heart, is devoted to authorial intentionality. No scholar wishes to be wrong about a text, or to be told that the meaning they found in it is invisible to others.

Ray Bradbury vehemently denied Fahrenheit 451 was about censorship. Although it is about books being burned, he did not intend it to be about book-burning as it is usually imagined; he thought he was writing about television setting us on a path toward ignorance of (and lack of interest in) great literature. His is hardly the last word: the author does not have sole interpretive power over their published work. But Bradbury’s opinion holds significant weight.

And that is why the unfinished work is so tempting. Its appearance is not what its author intended it to be, so authorial intentionality has less power. With an unfinished text, the reader is set free by the assumption of authorial unintentionality to try to divine what the text would have said if it had been completed. The interpretive possibilities may not be limitless — Shade’s poem Pale Fire is unlikely to be about Zembla, in the end — but they are certainly larger than with a finished, polished work.

In response to the doubts and rumors about her manuscript’s provenance, Lee released a statement: “I’m alive and kicking and happy as hell with the reactions to Watchman.” That she begins with “I’m alive” seems significant: nobody has been saying otherwise, exactly. But they have been writing about her manuscript as though it were that of a dead person.

Regardless of how Harper Lee feels about Watchman now, the Harper Lee who wrote the book fifty years ago didn’t want anyone to read it. She’s allowed to change her mind, of course, but the damage is done. The manuscript may have been rescued from a dusty archive instead of a pyre, but the effect is the same: when we read the book, we know we’re going against the wishes of the book’s author. And that knowledge gives us power.

But you knew all of this already. After all, you kept reading even after I said I wasn’t ready for this essay to be published. And that means

Donna Zuckerberg received her PhD in Classics from Princeton in 2014. She is the founding editor of Eidolon and teaches Greek drama at Stanford Continuing Studies and online for the Paideia Institute. Read more of her work here.

The views expressed in this article are those of the author and do not reflect the views of the Paideia Institute.