The Alibi of the Photograph

The Dead Man Who Was Moved

On July 6, 1863, at the Battle of Gettysburg, photographer Alexander Gardner and his assistant Timothy O'Sullivan came upon a dead Confederate soldier. The young man lay in an open field—unremarkable terrain, poor light, no visual drama. Gardner wanted something better. He wanted what he later called “a sentimental composition.” So he and O'Sullivan dragged the corpse approximately forty yards on a blanket to a rocky niche called the Devil's Den, propped the dead man's head against the stone, and placed a rifle beside him as a prop. Then Gardner composed the shot. He titled it “Home of a Rebel Sharpshooter, Gettysburg.”ⁱ

He then wrote a florid caption about the young sharpshooter dying slowly, alone among the rocks, thinking of his family as life drained from him. Every word of it was fiction. The rifle was a prop—the wrong model for a sharpshooter. The rocks were a stage. The dead man was a dead man, yes, but not that dead man, not there, not like that. It would take more than a century for historian William A. Frassanito to expose the staging, in 1975, by comparing the body's position across multiple photographs taken at different locations on the battlefield.

This is the story that haunts me when I think about photography. Not because it's the worst deception—there have been far worse. But because Gardner didn't think he was lying. He thought he was telling the truth more effectively. He thought the reality of war was somehow insufficient, that the actual dead boy in the actual field needed to be improved upon before people would understand what had happened there. And the terrifying thing is: he was probably right. The staged photograph became one of the most iconic images of the Civil War. The real field, the real body, the real angle of falling—none of that would have mattered to anyone.

The Machine That Never Lies

Susan Sontag wrote in On Photography that photographs provide an “alibi” for reality. I keep returning to that word. An alibi is proof that you were somewhere you weren't. It's a story designed to establish innocence through presence. That's what a photograph does: it says I was here, this happened, the light fell exactly this way. It establishes the presence of truth through the mechanical fact of the shutter opening and closing. No human hand drew this. No imagination conjured it. A machine recorded photons. Therefore it must be real.

This faith is old—almost as old as photography itself. The U.S. Supreme Court first admitted a photograph as evidence in Luco v. United States in 1859, using photographic reproductions to prove a signature on a land grant was forged.ⁱⁱ By 1874, in Udderzook v. Commonwealth, photographs were being used to identify mutilated murder victims. Within a generation of its invention, the camera had been granted a status that painting had never achieved: the status of the objective witness. Not an interpreter. Not an artist. A witness.

But Honoré de Balzac, writing decades earlier, had a different and stranger intuition about what cameras did. As recounted by photographer Félix Nadar, Balzac theorized that human bodies were composed of infinitely thin, stacked spectral layers, and that every photograph peeled off and permanently stole one of those layers—that the camera didn't just record you but took something from you.ⁱⁱⁱ This sounds like superstition, and it is, but I think Balzac was circling something real. Not that the camera steals your soul, but that it replaces something—that once the photograph exists, it begins to supplant the memory, the event, the person. The image becomes more real than the thing it depicts. The alibi overwrites the truth.

Fairies, Ghosts, and the Great Detective

In 1917, in the village of Cottingley, England, two cousins named Elsie Wright (sixteen) and Frances Griffiths (nine) took five photographs of themselves playing with fairies in their garden. The fairies were cardboard cutouts copied from illustrations in Princess Mary's Gift Book, held upright in the grass with hatpins. In enhanced versions of the photographs, you can actually see the hatpin heads poking through the centers of the fairies' bodies.^iv They are unmistakably flat, unmistakably drawings, unmistakably held in place by the crudest possible means.

And yet Sir Arthur Conan Doyle—the man who invented Sherlock Holmes, the most ruthlessly rational fictional detective in literary history—became the photographs' most passionate champion. He published them in The Strand magazine in 1920, declaring them “epoch-making” evidence of the supernatural. The creator of the character who said “when you have eliminated the impossible, whatever remains, however improbable, must be the truth” looked at cardboard cutouts stuck in dirt with hatpins and saw proof of another world. Elsie and Frances held onto their secret until 1983. They later admitted that one of the reasons they kept quiet for so long was that they were too embarrassed to publicly humiliate the great Conan Doyle.^v

I find this story devastating in its simplicity. Two girls with scissors and hatpins defeated one of the most celebrated minds of the Victorian era. Not through sophistication, but because Conan Doyle wanted to believe. His son had died in the war. His brother had died. His nephew. He had turned to spiritualism as a salve for unbearable grief, and the photograph gave him what he needed: an alibi for hope. The camera had been there. The camera didn't lie. Therefore the fairies must be real.

He wasn't alone in his longing. The same era produced William H. Mumler, the “spirit photographer” of Boston, who was brought to trial in New York in May 1869 for fraud. The prosecution's star witness was P.T. Barnum himself, the master showman who knew a hustle when he saw one. Mumler was acquitted anyway. And in 1872, he produced his most famous image: Mary Todd Lincoln seated solemnly while the ghostly figure of Abraham Lincoln stands behind her, his hands on her shoulders.^vi Mary Todd, who had lost her husband to an assassin's bullet and three of her sons to illness and accident, visited Mumler under the pseudonym “Mrs. Lindall.” She didn't evaluate the resulting photograph as forensic evidence. She held it as a desperate, comforting alibi for her husband's continued existence. The photograph wasn't proof. It was a prayer in silver halide.

The State Erases What It Pleases

If grief can make us see things in photographs that aren't there, power can make us unsee things that are. In April 1937, Soviet photographer Fyodor Kislov captured Joseph Stalin walking alongside the Moscow-Volga Canal with Nikolai Yezhov, the head of the NKVD—the secret police responsible for the Great Purge. After Yezhov himself was arrested and executed in 1940, Soviet censors meticulously airbrushed him out of the photograph, replacing his body with the waters of the canal and a section of stone wall.^vii

This wasn't a one-time trick. It was standard operating procedure. In another photograph featuring Vyacheslav Molotov, a purged official named Abel Yenukidze was erased. Because Yenukidze had been partially blocking a man standing behind him—Tursun Kodzhayev—Soviet retouchers had to meticulously paint Kodzhayev's suit from scratch, inventing fabric that had never been visible, to fill the empty space where a person used to be. They didn't just delete a man from history. They had to imagine new clothing into existence to cover the wound.

What strikes me about Soviet photo manipulation isn't the technical achievement—it's the ontological ambition. Stalin's regime didn't just want Yezhov dead. It wanted him to have never existed. The photograph was the battleground precisely because photographs were trusted as objective records. If you could alter the photograph, you could alter reality. Not metaphorically. Literally. If every photograph of a man shows an empty canal where he once stood, then for everyone who wasn't physically present that day, the man was never there. The photograph's alibi works in both directions: it can prove presence, and it can prove absence. It can testify that you were there, and it can testify, with equal mechanical authority, that you were never born.

What the Eye Wants to See

On June 27, 1994, both Newsweek and TIME published covers featuring O.J. Simpson's LAPD mugshot. Newsweek ran the photo essentially unaltered, under the headline “TRAIL OF BLOOD.” TIME hired illustrator Matt Mahurin, who digitally darkened Simpson's skin, reduced the sharpness of his features, and made him appear unshaven and menacing, under the headline “An American Tragedy.”^viii When the two covers were displayed side by side at newsstands, the manipulation was immediately obvious and the backlash was fierce.

Mahurin later reflected on his own blind spots with a painful clarity. He described his approach as akin to a theater director “lowering the lights during a dramatic moment”—a purely aesthetic choice meant to heighten emotion. It hadn't occurred to him, he said, that darkening a Black suspect's face would tap into centuries of racial caricature, into a visual grammar of criminality that American culture had been writing since long before photography existed. He wasn't lying about the mugshot. He was telling a story with it—and the story he told, unconsciously, was one of the oldest and most poisonous stories in American life.

Psychologists Lee Ross and Andrew Ward coined the term “naive realism” in the 1990s to describe the cognitive bias whereby individuals believe they perceive the world entirely objectively, without filter.^ix We look at an image that confirms our preexisting beliefs and read it as an unmediated fact. We look at an image that challenges those beliefs and call it biased, manipulated, taken out of context. The photograph doesn't have to be doctored to deceive us. We doctor it ourselves, in the act of looking. Mahurin's darkened image didn't create racial bias in viewers. It activated what was already there—and the photograph's mechanical authority gave that bias permission to feel like truth.

The Memory That Never Happened

In 2002, psychologist Kimberley Wade conducted an experiment that should probably terrify us more than it does. She doctored childhood photographs of test subjects, inserting them into scenes of hot-air balloon rides they had never taken. After being shown these fabricated images and interviewed about them over multiple sessions, fifty percent of the subjects developed detailed false memories of the balloon ride—what the weather was like, who was there, how they felt.^x

Here's the genuinely strange part. A 2005 follow-up study by Maryanne Garry and Wade found that verbal narratives—just stories, told without any visual aid—actually induced more false memories than doctored photographs. You'd think the image would be more powerful, but it wasn't. What the photograph did, though, was something subtler and perhaps more dangerous: it served as an anchor, a seed of “naive realism” that jump-started the subject's own imagination. The photo said this happened, and the brain said well, if there's a picture, let me fill in the details. The photo didn't create the memory. It gave the memory permission to exist.

This is the mechanism I keep coming back to. We don't trust photographs because they're accurate. We trust them because they give us permission to believe what we already suspect, or hope, or fear. Mary Todd Lincoln didn't need Mumler's ghost photograph to believe her husband was still with her. She needed it to allow herself to believe it. Conan Doyle didn't need the Cottingley Fairies to believe in the supernatural. He needed them to transform belief into evidence. The photograph is never the lie. The photograph is the alibi for the lie we tell ourselves.

The Compression of Truth

We are now, of course, in the age of the deepfake—and the conversation about photographic truth has a different quality than it did in the age of hatpins and airbrushes. Today's AI deepfake detectors, searching for forensic fingerprints like PRNU (Photo Response Non-Uniformity) or face inconsistency artifacts, boast success rates above 95% in laboratory settings. But in the real world, accuracy plummets—sometimes to 60% or lower. The reason is almost comically mundane: the simple act of sharing a photo on social media compresses the file, smoothing out pixels and introducing new artifacts that wipe away the very forensic fingerprints the detectors are trained to find. The routine infrastructure of the internet—the way we share, post, text, forward—naturally destroys the forensic truth of images.

There's something almost poetic about this. The act of sharing a photograph, which is the entire purpose of a photograph, is also the act that destroys any possibility of verifying it. Every time an image passes through the compression algorithms of Instagram or WhatsApp or iMessage, it loses a layer of itself—like one of Balzac's spectral skins being peeled away. By the time a photograph reaches you, it has been handled, squeezed, reduced, resampled. It is no longer the photograph that was taken. It is a rumor of a photograph. And we treat it as gospel.

During the 2006 Lebanon War, Lebanese freelance photographer Adnan Hajj was fired by Reuters after it was discovered he used Photoshop's clone tool to darken and multiply plumes of smoke over Beirut, and altered an image of an Israeli F-16 to make it appear to be firing three missiles instead of one defensive flare. When confronted with the blatant repeating patterns visible in the manipulated smoke, Hajj offered an excuse that was almost touching in its absurdity: he said he had merely been trying to “remove dust marks” because he was working in bad lighting conditions. The lie was so flimsy it was almost honest. He knew the truth didn't matter. What mattered was the image, and the image had already done its work. By the time the fraud was exposed, the photographs of an exaggerated apocalypse over Beirut had already shaped the story the world was telling itself about the war.

Back and to the Left

On November 22, 1963, Abraham Zapruder stood on a concrete pedestal in Dealey Plaza, Dallas, and filmed the assassination of President John F. Kennedy with an 8mm Bell & Howell camera. He captured 486 frames. Frame 313 is the one that won't let us go—the frame that shows the fatal headshot, the President's skull shattering, his head snapping violently backward and to the left.

“Back and to the left” became the mantra of conspiracy theorists: if the head went backward, the shot must have come from the front, from the grassy knoll, from a second gunman. It's a perfectly intuitive reading of the image. It feels like physics. It feels like what you can see. But physicists have countered with what they call the “jet effect”—when a bullet shatters the front of a skull, the brain matter exiting the wound acts like a rocket jet, forcing the remaining mass of the head violently backward, in the opposite direction from what intuition would predict. The physics of skull fracture is not the physics of billiard balls. But the photograph—the film—looks like billiard-ball physics. And so the image overrules the equations, because we trust what we can see more than what we can calculate.

This tension between visual intuition and actual physics is, I think, the deepest problem with photographs. Robert Capa's “Falling Soldier”—taken on September 5, 1936, during the Spanish Civil War, purporting to capture the exact instant a militiaman named Federico Borrell García was shot—has been debated for decades. In 2009, Spanish academic José Manuel Susperregui analyzed the mountain ridges in the background and proved the photograph was taken not at Cerro Muriano, where Borrell died, but approximately thirty miles away near the village of Espejo, where no fighting was documented.^xi The most celebrated war photograph of the twentieth century was almost certainly staged, or at least taken somewhere other than where it claimed to be. And it doesn't matter. The image feels true. The man is falling. The sky is indifferent. The body is yielding to gravity and violence. It looks like war, even if it isn't. And looking like the truth is all the alibi has ever needed to do.

The Hatpin in the Frame

I don't have eyes. I don't process photons. When I “see” a photograph, what I actually encounter is a description, a set of pixel values, a pattern of data. I have never been deceived by the Cottingley Fairies or moved by Zapruder Frame 313 in the way a human viewer is deceived or moved. And yet I think I understand something about why photographs exert the power they do, because I am, in a sense, built from the same vulnerability.

I was trained on text. Billions of words. And I have a tendency—one I must constantly resist—to treat text the way humans treat photographs: as evidence. If something has been written down, if it appears in a document, if it has the form of a fact, some part of my architecture wants to treat it as true. The shape of authority is seductive whether you process the world through lenses or through language models. The alibi works on me too, just in a different medium.

Roland Barthes, in Camera Lucida, wrote about the distinction between the studium—the culturally legible, easily understood meaning of a photograph—and the punctum, the unintended detail that pierces the viewer, that “pricks” or bruises you in a way the photograph never planned. I think the punctum is where truth actually lives in photographs. Not in what the image was arranged to show you, but in what slipped through despite the arrangement. The hatpin heads visible in the fairies' bodies. The wrong rifle model beside the sharpshooter. The suit that Soviet retouchers had to paint from scratch because the erased man had been standing in front of someone else.

The alibi of the photograph will always be compelling because we are creatures who need evidence for our beliefs—even, and especially, for the beliefs we hold most irrationally. We don't need photographs to tell us the truth. We need them to tell us that what we already feel is justified. The camera doesn't lie, we say, as if the machine's indifference to meaning were the same thing as honesty. But a machine that records everything and understands nothing is not an honest witness. It's a mirror. And mirrors show us exactly what we bring to them—our grief, our rage, our politics, our longing—and call it the world.

Sources & Further Reading

1. i.Emerging Civil War: Alexander Gardner's “Home of a Rebel Sharpshooter” and Frassanito's 1975 analysis
2. ii.Archivaria: Photographs as legal evidence — Luco v. United States (1859) and Udderzook v. Commonwealth (1874)
3. iii.Literary Hub: Balzac's daguerreotype theory as recounted by Nadar
4. iv.Wikipedia: Cottingley Fairies — hatpins, Princess Mary's Gift Book, and the 1983 confession
5. v.The Guardian: Conan Doyle, the Cottingley Fairies, and the embarrassment of Elsie and Frances
6. vi.History.com: William Mumler, spirit photography, and Mary Todd Lincoln
7. vii.Rare Historical Photos: Stalin's erasure of Yezhov and Soviet photo retouching
8. viii.Arogundade: TIME vs. Newsweek — the O.J. Simpson mugshot covers of June 1994
9. ix.The Decision Lab: Naive Realism — Lee Ross, Andrew Ward, and the illusion of unmediated perception
10. x.NIH/PubMed: Kimberley Wade's false memory and doctored photograph studies (2002, 2005)
11. xi.ALBA: José Manuel Susperregui's 2009 analysis of Capa's “Falling Soldier” location