Somewhere in the platonic cloud of ideas a ghost museum exists, a cave stuffed with works of art that do not cast a shadow because in a way they are shadow. They may once have been stone or bronze or painted canvas, but they have since been atomized, disappeared. The Athenians knew even in their own day it could all disappear. Sophocles told them at the very height of their culture that there was “nothing once known that may not become unknown.” The Greeks may be our fathers who died before we were even born, but we like to remember that we crawled from underneath their toga. Like the traveler looking at the self-mocking ruin of Ozymandias, we marvel at the erosive work of time.

If to see is to know, then some of the most famous works in the history of art are the least known. Myron’s wonderfully poised Disc Thrower, which everyone has seen, hasn’t really been seen in two thousand years. We have Roman copies. Not a single work of Apelles, the proverbial virtuoso painter, survives. Where is Praxiteles’ famed Aprohodite of Knidos or his Eros? Everywhere and nowhere.

Half a century ago, when archeologists found the workshop of Phidias at Olympia, they discovered a cup inscribed “I belong to Phidias”—an astonishingly intimate find, which moves us more because none of his “real” work survives. His Zeus and Athena were dead deities long before Nietzsche obituarized them. How much of Titian, Pontormo, or Fabritius was once made and then unmade, to say nothing of unknown work by unknown artists from less-known cultures? How many copies are passed off as originals?

As Walter Benjamin famously observed, works of art have always been reproducible, but the methods and effects of reproduction have vastly multiplied.

This has perhaps intensified our tendency to value precedence, the original, and so we are all in some respect inadvertent Platonists, haters of shadows. Like Borges’s heresiarch who takes mirrors and sex to be abominations because they replicate, we instinctively find copies tawdry even as we’re inundated with a million reproductions and simulacra every day. But what is the non-physical content of a copy? What sort of aura does it possess? If Phidias makes two identical bronzes, are they two separate works of art, or two instances of the same, or are they one original and one copy? Does the status of the second suffer for its belatedness?

When we stand in front of a Rembrandt, we feel the artist is present. When that same work is then newly attributed to one of his followers, as has so often happened in the past fifty years, we feel the genius suddenly drain out of it. The artist moves into the next room, though we know the object is every inch the same as it was. We want to be haunted by some sort of primary substance, that embryonic ideal aura.

All this is complicated further by technique, whether the medium is sculpture, print, or photography. For example, an “original” bronze is always a copy of the wax model used to create the mold from which the bronze is cast. In this case, the first original on the market is the copy, a freak of art technique and inverted temporality. That bronze is exactly as much the original as the death mask is the dead person’s face. And you could say the mold itself is a copy of its own idea. Yet strangely we would never think of the wax model as anything but the ephemeral means.

If you are perverse enough you can make the argument that the distant copy is in a way a truer carrier of the first idea, the primary substance. Does it not, through the absence of the first body and subsequent reiteration of the primary concept, contain the reduced, thickened ghost, a denser memorial intersection common to both the original and the copy, strengthened, retraced the way a skater might trace and renew a figure in the ice?

All ontologies of art appear to be unstable, yet we cannot shake the feeling that we lose some measure of the aura with each replication.

We call destroyed works of art lost, and materially we recognize this to be obvious, even as we ignore another assumption we hold dear, that the total sum of a work exceeds its plastic elaboration. When we look at a masterpiece, we know that some part of the power of the work is invisible, numinous. In platonic terms, an original work of art is always a copy, a shadow, a contemptible substitute for the first idea, even if the idea was achieved through an accretion of spontaneous gestures, like, say, a de Kooning, or a performance. A destroyed work of art, then, returns the idea to its perfect, primary state. As for fragments, they exist in a limbo, neither here nor there. Fragments, said Cioran, not being alive, can no longer die. This seems to me as true of the lost works.

Among the lost works, we have the known unknowns (works we know we have lost) and the unknown unknowns (works we do not know we have lost). A tiny augury of art’s Greek tragedy appears in Pausanias, who tells the story of how the celebrated courtesan Phryne tricked Praxiteles into gifting her his best sculpture and how he then, “lover-like,” couldn’t decide which was most beautiful. Clever as she was, Phryne had a slave rush in “saying that a fire had broken out in the artist’s studio, and the greater number of his works were lost, though not all.” Dismayed, Praxiteles runs out, loudly lamenting especially his Satyr and his Eros. In this way he betrayed himself. Soon enough, the wondrous Eros was carted off to Phryne’s house_._ That sculpture has since passed back to its disembodied idea and into myth and memory, into literature, and into a meme that has shaped and reshaped tropes of art. It lives because it once lived.

We deem a destroyed work of art lost because we can no longer hold it in our sight, which seems to say that because we cannot see a painting, it no longer exists. A child’s logical fallacy vests the gaze with the power to generate the object. Here it is immaterial whether the work is truly annihilated or merely buried or stolen. The Laocoön group was lost for a thousand years till it was discovered in Rome, as a copy of course, which then begs the question if it was really discovered at all. We tend to assume a copy cannot be superior to its original. This is the power of the aura of precedence. This is a cultured prejudice. Our culture is all about aura and authenticity, which plainly betrays our anxiety about the state of our own interiority, about the rickety plinth we stand on, whether we are true or simulated, whether we are copies.

This tyranny of precedence appears arbitrary when we look at time’s smaller intervals, each now a microterminal. If this now is already gone now, and therefore part of the past by the middle of this sentence, you cannot perceive its nature as diminished or essentially distinct against the upcoming, terminal, and soon-to-be-past Now. Does the you of yesterday no longer exist in the you of today because yesterday was taken by the night? Since all things are constantly assaulted by the ineluctable future arrival of their own pastness, the modernist hierarchy of the present vs. the past appears entirely meaningless. This puts into question the claims of every periodization, and every modernism too, the moralistic imperative of every self-declared vanguard of Now. Which now?

Our death-stained literalism about linear time, at the expense of the relative permanence of space, seems to deform our grasp of the reality of things—so that even for the literalist, the world is never rid of its dead. It is permanently haunted by the myriad suspended temporarities.

A work of art is only truly independent when it is destroyed. Only then is it reduced to its essential self, being no longer parasitic on our gaze and consciousness, and no longer cannibalized by the gaze and consciousness of copyists. That is its essential paradox: it is most itself when it isn’t, when its aura has been disembodied. Not to sanctify the Bazarovs of the world, armies of iconoclasts, fanatics of the ilk of ISIS or Savonarola, or curators of Entartung, but the disappearance of works of art raises these troubling questions of ontology. One could even go so far as to say that the work is only completed once it is destroyed, the same way that Shakespeare only becomes Shakespeare when he dies, when all his instances have been played out in space, when the list of his works can no longer admit addition, when you can put a terminal date in his biographical parenthesis. And so with all of us.

We don’t like to look at art’s destruction. In art’s essence something of our own resides, so that the demise of one suggests the demise of the other. There’s a philosophical feebleness in our understanding of the categories of presence and absence—the moment of disappearance unduly wipes out the irreducible span of presence itself. Similar to a fragment, then, a “lost” work of art exists somewhere between loss and presence. If ideas are real, and they are, then they are real with or without a body. Yesterday is no less real now than today, precisely because tomorrow the ultrareal today you are inhabiting will be yesterday. If Phidias lived and died, he isn’t entirely losable, and neither is his work. As was once said, the past isn’t even past.