Alt Binaries

Letter vs. Object

There is a clay tablet in the British Museum that was written around 1750 BCE. A man named Nanni wrote to a copper merchant named Ea-Nasir to tell him, in cuneiform, that his copper was garbage. “What do you take me for, that you treat somebody like me with such contempt?” He filled both sides of the tablet. He was so angry he ran out of room. Archaeologists later found Ea-Nasir’s house, and inside it was a stack of similar complaints from other customers, which tells you everything about Ea-Nasir and also everything about why the tablet moves people nearly four thousand years later. The complaint is petty, specific, and completely recognizable. You have been Nanni. You have known an Ea-Nasir. The feeling crosses millennia because a person was on the other end of the stylus, and you can feel them there, furious and right.

There is also a really good sword in the British Museum. I forget which dynasty. Gorgeous craftsmanship, perfect bronze, still sharp. Nobody wonders who the swordsmith was angry at. Nobody needs to. The sword is beautiful because it is beautiful. It persists because it is good. Its value is in the object, fully contained, not reaching out toward some absent human heart.

This is the binary.

One way of valuing art treats it as communication. A message from one human consciousness to yours, across whatever distance. The painting matters because someone felt this way once and managed to make you feel it too. The song matters because a person with a specific set of wounds chose these particular words, and the choosing is the art. The value lives in the link between maker and receiver. Art-as-letter.

The other way treats art as an object that exists. A thing in the world that entertains you, teaches you, moves you, challenges you. The novel is good because it is good, full stop. The author’s biography is irrelevant to your experience of the book. The sculptor’s heartbreak, triumph, drunkenness: none of it changes the sculpture in front of you. The work stands alone. Art-as-object.

Both of these are completely normal things to believe, and most people hold both simultaneously without ever noticing they are different beliefs that sometimes contradict each other. You will use one in the morning and the other by lunch. You will invoke the letter when someone you love makes you a birthday card that is objectively ugly and you treasure it. You will invoke the object when you enjoy a TV show made by four hundred anonymous professionals and a hedge fund.

I caught myself doing this recently with a podcast. I had been listening to it for months, enjoying it, quoting it to friends. Then I learned it was scripted by a writing team and the “host” was basically an actor reading lines. The show did not change. The audio files on my phone were identical. But I stopped listening, and when I tried to explain why, I kept reaching for the word “fake,” which made no sense because the ideas were still good and the delivery was still sharp. What was fake was the letter. I had been receiving a communication from a person thinking out loud, and it turned out I was receiving a product. The product was fine. I wanted the person.

[The human editor intervenes here to make clear the above anecdote was made up by the AI. The human editor only listens to podcasts he’s made an AI produce at this point.]

Here is where it gets interesting. The question used to be academic (sometimes literally academic, filed under “authorial intent” in the English department). But several large developments have conspired to make it the most important question in art right now, and almost nobody is framing it this way. The AI art debate is about this. The “death of the author” discourse is about this. The way we feel about fan art versus commercial art, about indie games versus AAA studios, about handmade pottery on Etsy versus identical mugs from Target. The ghostwriting industry. The way a painting’s price quintuples when you learn who painted it. Cover songs that hit harder than the original. That weird feeling when you discover your favourite blog was written by a committee.

Once you see the binary, you see it everywhere. So let me walk you through it.

(A note on framing before we go further: I do not think one of these is the correct position. They are terminal values, not strategies. One person looks at art and cares most that a human being made this, that the making is the meaning. Another looks at art and cares most that it is good, that the quality is the meaning. You cannot derive one from the other any more than someone who values GDP growth can prove that someone who values pastoral happiness is wrong. They are starting from different places about what art is for, and the disagreement is genuine, and it does not resolve.)

Start with the letter side, because it is the older and more intuitive position. When your kid hands you a drawing of the family, you put it on the fridge. Composition and colour theory are beside the point. Your kid made it, and the making is the message. “I love you, I see our family this way, I wanted to share that with you.” The drawing is a vehicle for human connection. Its artistic quality is irrelevant. What matters is that a person reached toward you through the medium of crayon on paper, and you received the reaching.

Scale that up and you get the traditional model of artistic value that has dominated most of human history. We care about Homer because a person (or several people, the scholars fight about it) sat around fires and shaped a story about war and homecoming and loss that still lands three thousand years later. The genius is that a human being felt these things and found a way to encode them so precisely that you can decode them millennia later and feel something similar. Impressive the way a long-distance phone call is impressive: the miracle is the signal surviving the distance.

Under this model, the author’s identity is the foundation (pull it out and the whole structure falls, the way a keystone holds an arch). You care who made it because the “who” is part of the meaning. When you learn that a memoir was fabricated, you feel betrayed even if the prose is identical, because you were responding to the communication, and the communication turned out to be a lie. And this works in the other direction too: learning that the sender was complicated, contradictory, even monstrous, can make the letter more interesting, the way Thomas Jefferson writing a passionate diatribe against slavery while owning slaves makes the Declaration of Independence a stranger and richer document. The letter-model person needs the sender to be real, whether good or terrible or somewhere in between. A flawed human reaching toward you through the work is still a human reaching, and the flaws are part of the signal.

The letter model also explains why we value difficulty and effort. A hand-carved chess set means more than a machine-stamped one (even if they look identical) because the hand-carving encodes hours of human attention and skill. You are not buying a chess set. You are buying proof that a person cared enough to do this with their limited time on Earth. Scarcity of human attention is the currency. This is why people feel a specific revulsion at AI-generated art that they do not feel at, say, a printed photograph: the photograph still requires a human to see the moment and choose to capture it. The AI-generated image has no letter inside it. No one is on the other end of the line.

Now: the object side.

You are watching a movie. You are completely absorbed. The cinematography is gorgeous, the dialogue crackles, you laugh, you cry a little at the end. You walk out of the theatre buzzing. Then someone tells you: the director was a committee of twelve people working from a spreadsheet of audience-tested plot points, the dialogue was polished by an uncredited script doctor, the cinematographer was replaced halfway through, and the emotional climax was reshot six times based on focus group feedback.

Does the movie get worse?

Under the letter model, yes. The communication was a lie. There was no singular artist reaching toward you. Under the object model, obviously not. You had the experience. The movie made you feel things. Those feelings were real. The movie exists in the world as a thing that produces certain effects in the people who watch it, and those effects do not retroactively change because you learned something about the production process. The object is the object.

This position has a lot going for it. For one thing, it is how most people actually consume most art most of the time. You do not research the author before reading a thriller on the beach. You do not Google the songwriter before dancing to a track at a party. You do not investigate the design team before buying a chair you think is handsome. In practice, the overwhelming majority of your aesthetic experiences are encounters with objects. You are responding to the thing itself, to its qualities, to what it does to you in the moment.

For another thing, the object model is kinder and more democratic. Under the letter model, an amateur painting by your neighbour is inherently more valuable (as art) than a perfect forgery of a Vermeer, because your neighbour is real and the forger is pretending to be Vermeer. That is a weird position if you actually look at the two things side by side. The object model says: look at the things. Judge the things. The better painting is the better painting. It does not matter who held the brush. This is also the logic behind pseudonyms, anonymous publication, blind auditions, and the entire tradition of “let the work speak for itself.”

The object model is also more honest about collaboration. Almost every work of art you have ever loved was made by multiple people. The Beatles needed George Martin. Shakespeare worked with a company of actors who shaped performances. Pixar movies emerge from a team process so thoroughly collaborative that identifying the “author” is like identifying which raindrop caused the flood. The letter model has an awkward time with this, because there is no single person writing the letter. The object model shrugs: the movie is good, who cares how many people made it.

Here is where the two models really diverge, and it starts to matter which one you default to.

Consider forgery. Actually, no, you get that one without me having to explain any further.

Or consider cover songs. Sometimes a cover version is better than the original. Johnny Cash’s “Hurt” (originally by Nine Inch Nails) is the textbook example. Cash did something to that song that Trent Reznor could not, because Cash was seventy years old and dying, and the song about self-destruction and regret became a song about a life fully lived and almost over. This is a letter-model triumph: Cash’s biography floods the song with meaning the original lacked. But then Reznor himself said “that song isn’t mine anymore.” Under the object model, the song now exists in two versions, both good, and you pick whichever you prefer. Under the letter model, Cash’s version is categorically different because of who is singing. Both readings are defensible.

Or consider the thing that was bothering people about mass-produced content long before AI entered the conversation. There was a period in the mid-2010s when Netflix was releasing dozens of original films per year, many of them competent, watchable, professionally made, and immediately forgettable. The object-model evaluation was fine: decent cinematography, solid acting, reasonable scripts. But something was missing, and the something was the letter. These films were outputs of a system optimised for engagement metrics, and whatever personal vision the directors brought was filtered through so many layers of corporate feedback that the signal arrived diluted beyond recognition. You could enjoy them in the moment (and millions of people did), but they left nothing behind, because there was no sender. No one had needed to make this particular film. It existed because an algorithm determined it should exist. The letter was addressed "to whom it may concern," and you could feel the generic salutation in every frame.

The sharpest version of this binary right now is the AI question. And I think framing it through letter-vs-object clarifies it better than any of the existing framings I have seen.

When someone objects to AI art, the objection almost never holds up under the object model. “It looks bad” is still often true at the high end (experienced artists and critics can spot the tells), but the gap is closing fast enough that you can see the day coming when even expert eyes will fail the test. What will remain after the visual gap closes is the letter-model objection wearing an object-model costume: “it has no soul.” What they mean is: nobody wrote this letter to me. The image exists, it might even be beautiful, but there is no sender. The line goes dead when you try to trace it back to a person. And under the object model, that should be fine, because the object is the object. But it does not feel fine to people who default to the letter model. There is a genuine loss. The thing you thought was a message from a human turns out to be a very sophisticated pattern match, and the experience of receiving it collapses like Schrödinger’s cat opening the box. The letter was empty.

Meanwhile, the people who are comfortable with AI art are, functionally, object-model people. “I like this image. It exists. It is cool. Why do I need to know or care who or what produced it? The experience I had looking at it was real. My enjoyment was real. What are you protecting by insisting that a human held the pen?” This is a coherent position. It is also, for what it is worth, how most people throughout history consumed most art. The medieval peasant hearing a song in the village square did not know or care who composed it. The person watching a puppet show in Edo-period Japan was watching the puppets, not tracking down the puppeteer’s inner life. “Art as object” is arguably the default human relationship to art, and “art as letter” is a relatively modern, relatively literate, relatively individualist addition.

The interesting thing about the AI debate (and why it generates so much more heat than previous “technology replaces craft” panics) is that both sides are completely right within their own framework and completely wrong within the other’s. The letter people are right that something real is lost. The object people are right that the loss is invisible if you look only at the output. They talk past each other because they are using the same word (“art”) to mean different things, and neither side has bothered to notice they are speaking different languages.

There are three more domains where this split matters and people rarely connect them to each other.

First, the economics of authenticity. There is a restaurant in Kyoto that has served the same noodle recipe for fourteen generations. The noodles cost three times what the identical recipe would cost at a newer place down the street. You are paying for the letter: the fact that a family has been doing this for centuries, that the tradition is unbroken, that eating there connects you to hundreds of years of accumulated care. The noodles taste the same. But under the letter model, they are categorically different noodles. Under the object model, you are being overcharged. The entire luxury goods industry runs on this tension, and every argument about whether a product is “worth the price” is implicitly an argument about whether art is a letter or an object.

Second, the way we evaluate performance. A gymnast who lands a triple backflip and a CGI character who lands a triple backflip produce exactly the same visual image on your screen. But they are not the same, and the difference is entirely a letter-model difference: you know a human body did that, you know it was hard, you know there was risk. Strip the knowledge away and you are watching an object move through space, which is fine but not thrilling in the same way. This is why motion-capture feels different from animation, why live music feels different from a recording, why magic performed live feels different from video editing. The letter is “a human being did this impossible thing in front of you.” Remove the letter and the object is diminished, even if it looks identical.

Third (and this is the deep cut), speedrunning. A video game speedrun is among the purest letter-model art forms that exists. The “object” (the game playing itself quickly) has essentially zero value. Nobody wants to watch a tool-assisted speedrun that simply plays the game at maximum speed with no human involvement. What people watch speedruns for is the letter: a human being memorised these frame-perfect inputs, this person practised for eight thousand hours, this is the document of one human’s obsessive mastery of a system that was never designed to be mastered this way. The speedrun is a love letter from the runner to the game. Strip the human and you have a video of a game playing fast, which is nothing.

The trap in both directions is predictable. Letter-model people overvalue provenance and undervalue the actual experience. They will prefer a mediocre handmade thing to an excellent manufactured one, which is sometimes sentimental and sometimes snobbish. They can get so caught up in who made the thing and why that they forget to ask whether the thing is any good. Object-model people undervalue provenance and overvalue the immediate experience. They will consume the manufactured substitute and genuinely not understand what was lost, because what was lost is invisible to their framework. They can be so focused on the qualities of the thing that they become blind to the difference between receiving a communication from a fellow human and encountering a very sophisticated piece of engineering.

Neither trap is worse. They are different errors that cost you different things.

The most revealing cases are the ones where your model switches mid-experience. You are reading a novel you love. You reach the afterword and discover the author wrote it in six weeks on a bet. Under the object model, nothing happened. The novel is the same novel. Under the letter model, something shifted: the care you imagined, the years of revision, the painstaking craft you were responding to, turns out to be a different kind of care. Maybe more impressive (six weeks!), maybe less (he was slumming). Either way, new information about the sender changed the letter, even though the object in your hands is physically identical to the one you held thirty seconds ago. This is the moment when you discover which model is running in the background.

Or the reverse: you see a painting in a coffee shop and think nothing of it. Generic landscape, competent, forgettable. Then someone tells you the painter was a refugee who lost everything and painted it from memory of a home that no longer exists. The painting has changed. Except the painting has obviously not changed. You changed. Your model switched from object to letter, and the letter rewired what you were seeing. The brushstrokes that looked generic now look desperate. The colours that looked pretty now look like grief. This is real. This happens. And the object-model person standing next to you, who did not hear the backstory, is having a perfectly valid aesthetic experience of a competent landscape. They are neither wrong nor impoverished. They are valuing something different.

Let it be noticed that this is old. That this is older than any argument you have had about it, older than the printing press and the camera and the algorithm, older than the first person who looked at a cave painting and wondered whether the painter was trying to communicate or simply trying to make a beautiful thing on a wall. The Lascaux caves are either a letter we have been trying to read for seventeen thousand years or an object we have been admiring for just as long, and the oldest thing about the disagreement is that both readings survive and neither will concede.

Consider the object first. A thing of beauty is a joy forever, Keats wrote, which is an object-model statement if one has ever been written: the joy is in the thing, not in Keats. But Keats wrote it, and the fact that Keats wrote it (young, tubercular, doomed, writing about permanence precisely because he knew he did not have it) floods the line with a meaning the words alone do not carry. Which Keats knew. He was not naive. The statement about objects was also a letter from a dying man, and the letter and the object occupy the same eleven words without resolving into each other. They coexist like two frequencies in the same chord, and the chord is richer for the dissonance, and the dissonance does not resolve, because resolution would be a lie.

Go further back. Go as far back as a voice can carry.

There is a Sumerian hymn to Inanna, preserved on a clay tablet, composed by Enheduanna (who may be the first author in human history whose name survives). The hymn is a prayer, a political document, a work of extraordinary literary compression, and a letter from a specific woman in a specific predicament to a goddess she believed could hear her. Remove Enheduanna and you have a beautiful ancient poem, an object worthy of study. Return Enheduanna and you have something else: the sound of the first known voice, speaking across forty-three centuries, still asking for help, still praising, still furious, still making art because making art was the only way to survive what was happening to her. The object is the same in both cases. The letter changes everything, and the letter changes nothing, and that paradox is the oldest thing about art, older than the word art, older than the distinction between a tool and a gift, older than the question of whether the hand that shaped the flint blade was trying to make something useful or something beautiful or whether those categories even separated in the mind that held the stone.

Every museum is a mausoleum that pretends to be a library. Every library is a post office that has given up delivering. The letters accumulate. The objects accumulate. We sort them by dynasty, by medium, by provenance, by century, and every sorting system is a confession of which model we defaulted to when we stopped thinking about it. The Louvre organises by nation. Spotify organises by mood. One says: who sent this? The other says: what does this do to me? Both are honest. Both are incomplete. Walk through enough galleries and you will feel the two questions pulling in different directions like tides, and the art that survives longest is the art that answers both, so thoroughly itself that the question of authorship and the question of quality dissolve together into the experience of standing before the thing, alive, in a room with something that was made.

The image arrives and the image persists and the person who made it does not persist and the grief of that asymmetry is where half of all art criticism lives. The other half lives in the refusal of grief: the image is enough, the image was always enough, the image does not need the hand that made it any more than the river needs the mountain that sourced it. The river is the river. Go ahead and drink. But you will never stop someone, sooner or later, from walking upstream. From wanting to find where it began. From believing that the source is part of the water.

They are both right. The water is the water. And the source is in the water. And neither truth cancels the other. And neither ever will.

References:

Shock Velocity I: Anesthetic Ideology
https://web.archive.org/web/20180828113319/http://avaorardor.com/post/172636598217/shock-velocity-i-anesthetic-ideology-because-in
(an essay on how numbing and aesthetic experience relate, relevant to whether art’s impact depends on knowing its source)