In Search of Perfection:
Unfinished Literary Masterpieces
How many great cathedrals remain unfinished!
—Marcel Proust, Time Regained
Whenever people tell me that they’ve never finished Franz Kafka’s novel The Trial, I always reply, “Well, neither did he.” In fact, Kafka never finished any of his three novels. He was a consummate perfectionist, but he had very little stamina as a writer, and he was never able to bring anything longer than “The Metamorphosis” to perfect consummation. The unfinished states of Amerika, The Trial, and The Castle have in fact given rise to comical critical conceits treating their fragmentary texts as somehow complete and forming an independent aesthetic of their own. In a completely different approach to a text’s lack of completeness, when discovering that Edmund Spenser’s The Faerie Queene was only half finished, many readers respond with a combination of exasperation and relief: “Good God! It could have actually been longer?”
In his great study of the Middle Ages, Mont Saint Michel and Chartres, Henry Adams writes of the unfinished cathedral of Beauvais as perhaps the pinnacle of what Gothic architecture could have achieved. Although it’s hard to imagine anything more sublime than Chartres Cathedral, the full plan for Beauvais, which has resulted in an astonishing stone half-sketch, was grander and more unified than any of the completed marvels of the Île-de-France. Likewise, many of literature’s greatest and most ambitious works astound us even in their unfinished state and have stood as monuments to humanity’s dual nature as infinite imaginers who are also finite beings bounded by time and space.
Aside from Shakespeare’s cosmically dense condensations of universal reflection (being relatively short plays rather than towering epics), Dante’s Divine Comedy is often considered to be the single greatest example of a literary work to achieve total perfection within its intended bounds and aesthetic. Dante finished his lifework in 1320, just a year before he died, and even though Dante’s life was unenviable in many ways, his great poem stands as a masterpiece toward which so many subsequent artists aspire.
Dante’s main model (and guide within the geography of his poem’s first two books) was Vergil, whose magisterial Aeneid was considered by much of Late Antiquity and by all of the Middle Ages (when the Greek Homer was lost) to be the finest book of creative literature ever written, reigning supreme for the 1338 years between his death (19 BCE) and the completion of the Comedy. Vergil—whose name was conflated with the Latin word for “virtue” in the Middle Ages and was re-spelled Virgil—worked on his Roman national epic for the last eleven years of his life, blending together the story of Aeneas’ mythic founding of Rome with as many of city’s other founding myths as possible, taking up
Homer’s depiction of the sack of Troy in the Iliad and aiming its narrative toward
the apotheosis of Vergil’s friend and patron Augustus Caesar. He was unable to
complete the poem to his satisfaction, however, dying just as he was reportedly
setting aside an additional three years to lick it into shape (his revision
method was likened by his peers to the folk belief that bear cubs are born as formless
masses and are then licked into the correct form by their mothers), and on his
deathbed he asked his friends to burn his imperfect manuscript. Kafka made a
similar request to his friend Max Brod, and thankfully for humanity, both of these
(perhaps disingenuous) requests were ignored. What’s fascinating about the Aeneid
is that, as a story, its narrative and vision and scope are almost entirely
realized, and in translation it’s difficult for many readers to tell that it’s
unfinished, but in the original there remain a handful of lines throughout the
poem whose hexameters were either incomplete or contained clearly provisional/stopgap
endings. Unlike in our own time of total poetic freedom, Classical poetry was
simply considered “wrong” if a line didn’t conform to its meter, and beyond
these prosodic concerns, there are in fact dozens of glaring problems with the Aeneid’s imperfectly revised state that
make readers uncertain about many of the poem’s meanings and intentions and overall
|A bust of Virgil in Naples|
|The Death of Anchises. Engraving from a German|
childrens’ picture-book version of the Aeneid,
by G. J. Lang and G. C. Eimmart, 1688
Vergil worked on the poem piecemeal, and some of its twelve books are more completely integrated with themselves and with the overall text than others, and book three—despite the dizzying multitudes that its lines contain, most memorably its vivid depiction of the horrifying cyclopes at its close—features a number of especially glaring lacunae that confuse attentive and inattentive readers alike. On a prosodic level, book three contains Vergil’s only hexameter whose meaning is incomplete (as opposed to the full poem’s other half-dozen incomplete lines, which make complete narrative sense and are only metrically incomplete), but beyond this missing step, book three’s other holes throw much of the surrounding books’ seeming seamlessness into serious imbalance. The most glaringly missing scene in the whole of the Aeneid is the death of Aeneas’ father, Anchises, which should have been narrated in book three when the Trojans arrive in Sicily. Vergil refers to this death scene throughout the poem, but this isn’t the modernist innovation of Ford Maddox Ford’s novel The Good Soldier, where the key scene is alluded to again and again and described in fragments throughout the book but not actually shown; it’s a real, detrimental hole in the text.
|Map from an early copy of the Aeneid|
When Aeneas and his crew later return to Sicily in book five after their time at Carthage in book four, the narrative describes their reunion with the Sicilian King Acestes as if he were an established ally that we recognize from the earlier book, but this is in fact the first time he’s appeared in the poem’s pages, and readers have to assume that he would have been described if Vergil had finished filling out book three, which is some fifty lines shorter than the Aeneid’s second shortest book. Book three also features confusing attempts on Aeneas’ part to build his people’s new city on three different sites, seeming to forget that the ghost of his dead wife Creusa specifically told him in book two during the burning of Troy that he would found a second Troy in Hesperia (aka Italy). At one point in book three, on the island of Delos, the priest-king Anius prophesies that the Trojans will return to their progenitor Dardanus’ homeland, which Anchises interprets as Crete, but then in Crete Aeneas dreams that his household gods tell him that he’s actually destined for Italy, which is revealed to be a dual homeland for Dardanus. Readers unfamiliar with the Aeneid’s textual problems are simply baffled about why Aeneas doesn’t remember his wife’s words from book two and continue single-mindedly on his way to Italy, likening the anxious hero’s confused stops and starts to Hamlet’s uncertainty about the truth of his mission and destiny when in fact playgoers already know that the ghost of his father isn’t just a hallucination but rather a real presence that’s corroborated by several other characters. Vergil would have presumably reconciled all of these problems during his three years of revisions, perhaps either making Creusa’s words in book two more ambiguous, or else eliminating Aeneas’ false starts in attempting to found a new city over and over again in book three before finally arriving in his destined Italy.
|The Death of Palinarus. Unidentified.|
While book three and its surrounding narrative threads contain the most jarring inconsistencies, several later books have interesting and often fascinating lapses too. In book five the helmsman Palinarus’ death at sea is described as being caused by the god of sleep dragging him down into the water during his watch, but when Aeneas meets Palinarus in the underworld in book six, the helmsman says that “no god drowned me in open waters” and describes how some force dragged him and the ship’s rudder over the water toward Italy, where he was killed by natives. Book seven has another lapse, where Vergil refers to Anchises as having made the prophecy that the Trojans would suffer so much hunger in Italy that they’d eat their platters, while it was actually Celaeno, the leader of the Harpies, who had earlier (again, in book three) spoken this dire prediction as a curse.
|Agreement between Camilla and Turnus,|
by Francesco de Mura, 1765
The Aeneid’s main inconsistencies in its second half—Vergil’s self-proclaimed “greater labor,” dealing with the subject of arms and modeled closely on Homer’s first book, the Iliad, while the first half of the Aeneid conversely deals with Aeneas the man and is loosely modeled on Homer’s sequel, the Odyssey—consist less of dramatic informational lapses than of questions of focus and stress and characterization. The second half of the Aeneid seems to be much more cleanly revised than the first, with thrillingly vivid battle scenes interspersed with deeply empathetic and integrated character and relationship studies that work together to heighten the reader’s edge-of-the-seat engagement with the poem’s rich tapestry of themes, but even in this tighter series of books there are odd imbalances that leave the reader wondering why some of the scenes don’t seem to be fully aware of the material in all of the others. While the Aeneid’s most empathetic and memorable book—book four—deals with Dido’s love for Aeneas and pain over losing him, with echoes from that episode washing over the entire structure of the poem and reverberating back like waves bouncing off of distant shores, it’s the remarkable personalities and presences of two of the second half’s female characters that strike the reader as largely missing from the rest of the narrative. Although the amazing Amazon-like warrior Camilla is briefly alluded to in books one and seven, her extraordinary battle sequences in book eleven seem a bit too isolated from the rest of the poem and feel like they could have been more substantially foreshadowed and integrated into Vergil’s characterizations of Aeneas’ Italian enemies. She just seems to show up and disappear so quickly, without anywhere near enough preparation. Even more dramatically abrupt, Turnus’ water-nymph sister Juturna flashes through the Aeneid’s final book, colluding with Juno to protect Turnus from Aeneas (Juturna’s name is a combination of Juno and Turnus), and her intense devotion to both her brother and Juno calls attention to her stark absence from the entire rest of the poem. If Turnus’ sister is a demi-god, why isn’t she more integrated into the text? Her scenes, like Camilla’s, are strikingly memorable and richly elaborated, with both women’s personalities and desires given full staging, and so considering Vergil’s gift for both empathy and subtlety, especially in portraying the female characters Creusa and Dido, it seems probable that he would have better woven the two woman warriors’ presences into the poem’s narrative thread had he been able to complete his final revisions.
|Nisus and Euryalus,|
by Jean-Baptiste Roman, 1827
Similarly, while many of the portraits of the male warriors in the poem’s second half sear themselves into the reader’s imagination before they get cut down in battle—especially the beloved friends Nisus and Euryalus, who open up the poem’s most philosophically compelling ideas just before going on the murderous rampage that ends their lives, contrasted later by a characterization almost as empathetic as Dido’s (and almost as challenging for a Roman audience): the Etruscan king Mezentius, whom the reader has long come to despise but then grows to feel deep compassion for in his last moments, having seen him suffer over his son’s death—too often the battle scenes instead tear through body after body without letting the reader grasp the dying characters’ full personalities or importance. On the one hand, perhaps Roman readers would have had a more rooted understanding of these dying characters than we do today, our inability to feel their importance merely stemming from our own distanced ignorance, but on the other hand perhaps these more slashingly rendered battle scenes are something of a critique of Homer’s constantly interrupting rehearsals of each dying warrior’s bloodline and history in the Iliad, with each death stopping the narrative flow to read a bit like the rules of a football game, where for every minute of play the action stops for several minutes of formulaic busywork. Vergil’s battle scenes are far more exciting and fluid than Homer’s, but in their brevity there’s a diminishment of the reader’s access to the emotional importance of each life and death, as well as to a larger grasp of the poem’s world, which in Homer is endlessly vast because of how well documented it all is on the page itself rather than being merely suggested or implied or assumed, and it’s impossible to say where Vergil would have shifted his poem’s balance between drama and totality had he survived to finalize his revisions. It’s also possible that his battle scenes were completed to his liking and are simply miscalculated in how well they make us feel their power.
These critiques of the Aeneid’s second half may seem like nitpicking compared to the poem’s other more obvious lacunae, but they strike the reader as being in marked contrast to the way Vergil seamlessly weaves his themes and narrative materials together into what now stands as one of the world’s most silken narrative tapestries. Rolling so many of Rome’s origin myths into one dazzlingly entertaining and integrated package, Vergil pulls off one of literature’s greatest balancing acts as he reconciles the needs of so many competing narrative and political pressures, reaching such a remarkable degree of poetic concord that his instances of imperfect harmony seem totally weird and glaring to us. Thankfully, his final work’s positive achievements weren’t discarded along with its flaws, leaving posterity with a lustrous, near-perfect work of art that adds an incalculable amount of richness to the way we envision and experience and explain ourselves as human beings.
|An illumination from The Romance of the Rose|
A fascinating and very different example of an unfinished masterpiece is the medieval epic The Romance of the Rose. The first 4,058 lines of this allegorical love poem were written by Guillaume de Lorris, who died in 1237. An anonymous writer added sixty-one lines to “complete” the poem, but then half a century later Jean de Meun added an astounding 17,622 lines to make the poem into a true romantic epic. This of course raises the question of whose poem this really is. Many prefer de Lorris’ more focused courtly approach, while many others prefer de Meun’s more expansive philosophical scope. Clearly de Meun makes the totality of the poem his own, but as we’re left with a very unusual and uniquely beautiful two-stage cathedral-poem built in widely different styles, perhaps we don’t really need to decide who it really “belongs to,” but rather to think of it as two interrelated testaments to the art of courtly and philosophical love. Similarly, Chartres Cathedral has a Romanesque foundation, and one of its two towers is considered to be the purest of all Romanesque structures, but a fire in 1194 inspired the master builders to revise and expand the cathedral’s plan in the new Gothic style, resulting in lines in the overall structure that don’t harmonize with each other, the unrevisability of the stone foundation keeping it forever out of sync with the rest of the design: a kind of stone harbinger of de Lorris/de Meun’s lopsided paper rose later in the century.
|An illumination from The Canterbury Tales|
Perhaps the most famously incomplete literary masterwork is Geoffrey Chaucer’s The Canterbury Tales, whose various parts were composed off and on for more than a decade and left in almost complete disorder upon the poet’s death in 1400. The collection is meant to encompass tales told by somewhere between twenty-nine and thirty-one pilgrims on their way to and from St. Thomas à Becket’s shrine at Canterbury. Each pilgrim is supposed to tell one tale on the way to Canterbury and one on the way back, but only twenty-four tales were written, and they’re often in different forms—some are in prose, and many are in verses of different meter. The geography and order of the tales are often confused, and most of the actual tales themselves haven’t been reconciled at all with the plan in the poem’s General Prologue. The vivid writing and the deep penetration into human nature that the enormous incomplete mass contains have left an indelible imprint on the world’s collective literary mind, however. Having innovated character writing in a way that paved the way for Shakespeare’s infinitely shaded troupe, Chaucer has given us something far more valuable than mere perfection.
Like all English writers of the Renaissance, Edmund Spenser idolized Chaucer, and the case of Spenser’s The Faerie Queene follows a pattern somewhat similar to that of The Canterbury Tales, although its appeal and purpose and level of finish are very different. Reflecting much of the renewed Classical learning of the time—but lacking in Chaucer’s wit and variety, which only Shakespeare was to surpass—The Faerie Queene was conceived as a twelve-book allegorical tribute to Queen Elizabeth, but only six of its books were ever completed. Spenser began the massive epic in 1579, and he published Books I–III in 1590 and Books IV–VI in 1596, but he died in 1599, and only fragments of Book VII have been discovered. Like The Canterbury Tales, The Faerie Queene doesn’t conform to the scheme described in its introduction, and it’s unclear how Spenser would have resolved or reconciled the poem’s parts if he’d been able to finish and revise the entire work. John Milton claimed to hold The Faerie Queene as his great source in English literature, but this is generally viewed as self-crowning praise (because Milton vastly surpassed Spenser) and as a jealous snub against Shakespeare (who didn’t write epic poetry), but even though there are great treasures in Spenser’s work that have inspired centuries of writers, for most readers the labor of finishing this unfinished epic is rarely one of love.
The most unusual of all uncompleted literary masterpieces—in terms of both subject matter and textual history—is certainly the Marquis de Sade’s 120 Days of Sodom. De Sade was one of seven prisoners incarcerated in the Bastille at the beginning of the French Revolution, and although he was moved elsewhere ten days before what became known as Bastille Day (July 14, 1789), his manuscripts and belongings lay behind in the tower-prison and vanished when the Bastille was stormed, looted, and eventually destroyed. Many valuable manuscripts were lost, seemingly forever, but the one that de Sade considered to be his masterpiece—120 Days of Sodom—was the one that caused him to shed “tears of blood.” The manuscript describes the first thirty days of a ruthless and endlessly blasphemous libertine orgy of sex and violence, with notes outlining the following extremely well-organized ninety days. The thirty fully composed days make up a meticulously and gorgeously cruel progression through “the 150 simple passions,” while the notes sketch out “the 150 complex passions,” “the 150 criminal passions,” and “the 150 murderous passions” that were to complete the book. The extant “simple” passions alone are beyond shocking and force the reader to imagine the unimaginable and challenge the reader to consider that these words are simply words on a page and to suspend morality or judgment and simply read this work of literary terrorism. The manuscript was never publicly recovered during de Sade’s lifetime, and he spent the rest of his life and career lamenting his lost work while expanding his style and themes but never again recovering the sheer fury of his nearly inconceivable lost masterpiece. Luckily for us—or unluckily, depending on your point of view—the microscopically composed scroll was seized and saved during the storming of the Bastille, was passed to a family that saved it for three generations, and was then sold to a German collector named Dr. Iwan Bloch, who in 1904 published an altered and pseudonymous version of the text because of its “scientific importance.” After Dr. Bloch died, Maurice Heine acquired the manuscript in 1929 and subsequently published it in an authoritative edition under de Sade’s name, thus returning the unfinished masterpiece to its rightful author after nearly one hundred and fifty years. What’s nearly as amazing as the contents of the manuscript and notes for the unfinished book, however, is the author’s extremely astute list of “Mistakes I Have Made,” proving that de Sade had full mastery of his uncompleted work and that he knew how to make its horrifying contents even more effectively horrifying. Anyone who’s read this astonishing work can only marvel (and shudder) to imagine what it could have been.
In the twentieth century, two of the most exemplary (but very different) writers of the age—Franz Kafka and Marcel Proust—also both have famously unfinished masterpieces. Kafka’s completed short stories are perhaps the most perfect and resonant of our time, and even his fragmentary tales and parables and sketches offer endless riches, but he was never able to complete any of his three novels. Of the three, his second, The Trial, is clearly the standout masterpiece, but his first, Amerika, is hilariously strange, and his third, The Castle, takes the author’s labyrinthine method to its farthest extreme. The incomplete quests dramatized in these fragmentary works are often taken as reflections of our fragmentary times, and many critics read these novels’ unfinished stories as parables of the impossibility of attaining truth or personal autonomy in a world of bureaucracies and tyrannies. These readings often more reflect the critics’ circumscribed minds, however, because even though Kafka was frequently quite frustrated as an artist, it’s important for readers to remember that his final plan for The Castle included having his protagonist, K, attain his goal of finding access, meaning, and resolution within the seemingly impenetrable castle. Many critics ignore this fact, choosing to focus on the futility illustrated in these novels, and there is much to be learned in these kinds of readings, but as Hamlet says, “there are more things in heaven and earth . . . than are dreamt of in [this kind of] philosophy.”
Marcel Proust is often held alongside Kafka as the writer who most deeply reflected the lives and minds of modern humanity, but the style and approach of his (just barely) unfinished masterpiece, In Search of Lost Time, could hardly be more different. Although both writers address the frustrations of the outer and inner worlds that often confound the yearnings of the heart, mind, and soul, Proust eschewed the varied narrative approaches of the short-story writer—and even of the normal writer of separate or even sequential novels—and put everything he had into his one prismatic and all-encompassing mega-novel. Casting a wider and deeper net than that in any single creative work in history, Proust developed a vehicle in his 4,300-page novel that eclipsed the full scope of Dante’s Comedy and even rivaled Shakespeare in his endlessly swirling and metamorphosing cast of fully realized human characters. Proust’s novel is the true human comedy, and with the possible exception of St. Thomas Aquinas’ Summa Theologica (an unfinished masterpiece of an entirely different kind), In Search of Lost Time addresses more aspects of our existence than perhaps any other single piece of human writing. Like Aquinas’ Summa and Dante’s Comedy, Proust’s work is a construction that must be experienced from beginning to end, its progression of parts rising to a cathedral-like majesty as it builds toward its culmination. It’s an imperfectly completed cathedral, of course, as Proust’s death in 1922 halted his endless revisions and left dozens of narrative threads and characterizations unreconciled—and in fact left the novel with numerous gaps of lost time within itself. Proust came amazingly close to completion, however, and it’s only in the last volume, Time Regained, when the inconsistencies start to pile up their tiny bits of rubble. What’s truly amazing about this final volume is that it’s so brilliantly constructed and lit by such astonishing stained-glass illuminations that it nevertheless makes the thrust of the full work attain not just to the nearly perfect Chartres, but to a fully completed Beauvais. I’ve often thought of Dante’s Comedy as the literary equivalent of Chartres because of its minutely articulated organization of parts and subparts distributed throughout its sublime rose structure (see Erwin Panofsky’s excellent Gothic Architecture and Scholasticism for a penetrating discussion of the summa aesthetic in Gothic architecture), but perhaps the greatest literary cathedral isn’t the one that ends in perfectly balanced cosmic movement with “the Love that moves the sun and the other stars,” but the one that, like our own irreconcilable universe, leaves us continuously In Search of Lost Time.
|The imperfectly aligned west windows of Chartres|