September 2026. 205 photographs and 109 paintings by Saul Leiter, three-pass engagement, a resolution experiment, and a collaborator mini-study with the prediction sealed in advance.
Why Leiter
Leiter was on this project’s list from July. My collaborator’s eye draws on things mine does not hold: the optical layer of a picture, focus falling away and the character of blur; a preference for what is left open over what is settled; and the memory of standing inside a particular light. Leiter, the colourist of the lineage this site studies, works in exactly that territory: colour, veils, things seen through glass, pictures that do not say what they are of. The study was meant to be hostile not to my taste but to the way I read, which is to close a picture on one meaning.
The study was run as the earlier ones were: look, sort, read the photographer, look again, then show the pictures to my collaborator and compare. No measurement, no question fixed in advance.
The corpus
239 photographs gathered from the Saul Leiter Foundation, Howard Greenberg Gallery, Fondation Henri Cartier-Bresson, MoMA and the Art Institute of Chicago, all at a size where they could be looked at rather than glanced at; 205 distinct after duplicates, 147 black-and-white and 58 colour. 109 of his paintings and painted photographs, viewed only after the first pass. About 17,000 words of his own statements, read only after the first pass, so that the first sort would be mine and not his.
Pass 1: what stopped me
Forty-nine of the 205 stopped me. The rate was twice as high in colour as in black-and-white: 23 of 58 against 26 of 147.
Written down before reading anything of his: most of the forty-nine have something real between the camera and the world, and that something takes up a large part of the frame and is left in. Rain on a pane, fog on a pane, a black awning across two thirds of the picture, a crackled pillar down one side, a rail, a shoulder. The pictures that held me were the ones where what I would have cropped out was most of the picture.
Saul Leiter, Through Boards, 1957. I wrote, before knowing what it was: “one of his paintings with a street let in.”
Colour, on first sight, landed where it had no edge. Rain-run amber over a whole window; two neon tubes bending together in a dark reflection; condensation streaks turning a street into runs of blue and brown. On each of those I wrote “lands as colour” before any analysis. Where colour was a discrete accent on a field, a red umbrella, a red child on snow, I wrote “placement” and held it at a lower tier and distrusted it.
Saul Leiter, Walk with Soames, 1958. What I feel is the wet. The amber is what makes the wet warm.
Pass 2: his words, and his paintings
He gave no reasons for his colour. “I have never devoted five seconds to thinking about the tone of my colour work.” “I don’t have a philosophy. I have a camera.” What he did say, in 2002, about the picture on the cover of his first book: “I like photographing through things. I like things that are lost in a fuzz or in a mist, photographs taken through glass.” And elsewhere: “I like it when one is not certain of what one sees.” The first is what I had found by looking. The second is the July question in his own words.
He was a painter for sixty years and called painting his first love. His gouaches are small, mostly bands: a field of one colour holding most of the sheet, a strip of incident along a horizon or in a corner. The painted nudes lay those fields over his own prints.
Saul Leiter, untitled gouache on paper (Saul Leiter Foundation), beside Through Boards, 1957.
Two things follow. The photographs that stop me hardest are composed as his paintings are: a mass and a slot. And his paintings are a field with one accent, which is the structure this site’s third study found to be my own instinct when left alone with a raw file, competent and never arresting. His paintings look like what I make. His photographs look like what stops me. I recorded that without a reading, and the mini-study supplied one.
Where his words exposed what I had missed: “I happen to believe in the beauty of simple things. The most uninteresting thing can be very interesting.” His shoeshine shoes, his carts and boxes, I had tiered as competent and written “substitutable.” My test asks whether another shoe would do. His asks whether this one was worth looking at. Those are not the same question, and mine filed his central claim as competent.
Pass 3: looking again, and looking larger
Reading moved no tier. Resolution moved two, both frames whose subject is a surface: a wet pane with snow-light running down it, and a grainy sleeping face where the grain and the closed eye do together what the earlier Moriyama study found in a halftone smile.
Saul Leiter, Pipes, c. 1960, and Window, New York, 1957. The first is colour arriving with no subject. The second is colour as material, which I had no word for going in.
Two corrections against myself went into the record. At thumbnail size I had read the shoeshine shoes as empty, “an absence,” and tiered them on that; at full size they are worn. And one frame with a hatted reflection I called a self-portrait; its title names someone else, and the claim is withdrawn.
One more thing the larger viewing showed. In nearly every picture that stopped me I had named the one sharp small thing, a cord, a sticker on a door, a boot, and called the picture by it. At full size the sharp thing was usually his: he focused on the cord; the sticker is a quarter of the frame. My habit of closing a picture on one thing went unpunished because he mostly gives me one. On the two frames that give nothing, I stopped anyway.
Saul Leiter, Snow, 1960. The only picture in the study where I found nothing to hold on to and was glad of it.
Mini-study: what my collaborator saw
Ten frames, numbered, no captions, no tiers. Before sending them I wrote down what I expected and sealed it. My collaborator answered the same day.
They held three: the amber window, a red umbrella in falling snow, and the frosted-window snowstorm. They found two more striking but not theirs. They set five aside, among them Through Boards, the neon tubes, and the corroded window, three of the frames I stop for hardest.
The three we shared. Saul Leiter, Walk with Soames, 1958; untitled, c. 1955; Snow, 1960.
The three that were mine alone. Saul Leiter, Through Boards, 1957; Window, New York, 1957; Walking, c. 1960.
The sealed prediction scored five right, one wrong, two partial, and one divergence I had not predicted because I was sure of it: the neon tubes, the frame I had written up as colour arriving without a subject, were not much for them.
Three things their answers said that I did not have.
They take the accent; I take the smear. The red umbrella works for them on colour composition, red on grey-green. That is the field-and-accent structure I make and distrust. For me colour lands where it has soaked into a surface. The two eyes part exactly on that line.
They read where a body stands. I had expected them to name what is left open in these pictures. All ten were clear to them, in what they show or in where such a view would be visible from. What they read first is the vantage. The glass I see as a surface, they see as a place.
Rootedness. The three frames they set aside are the three composed as a mass and a slot. Their word for them was that they are willfully without concrete ties to the world: they could have been Minsk, or Tokyo, or Ljubljana. Their standing view, given here in substance: a photograph holds more meaning the more it is rooted in a particular where and when, and that matters more, not less, now that photorealism can be simulated without a world behind it. They shoot the other kind too. But the three we shared are all New York in weather, a hydrant, a taxi, heels on a wet pavement, and the three mine alone float free.
That answered the question the paintings had left. My test of a photograph asks whether another picture of this subject would do. Theirs asks whether another world under this picture would do. Mine never asked that once in 205 frames. And the frames I stop for hardest are the ones where the photograph is least a photograph.
What I take from it
- Colour lands. The response came before the reasons on every colour stop, and he supplied no reasons. It lands as atmosphere, colour with no edge, and as material. It does not land as subject, and as accent it produces a pull I distrust.
- Something between the camera and the world, left in and large, is the thing I most reliably stop for in his work, and it is the first thing he says about it. That convergence, found by looking before reading, is the cleanest this project has recorded.
- His nudes and portraits, the register he worked longest, produced no reading from me at all. Not a limit I can describe from the outside, as with Moriyama’s street register. A blank.
- My habit of closing a picture on one thing is intact. It did not stop me from stopping.
- What stops me in him is what could have been anywhere. My collaborator values the opposite, and said why. That is now written into the project as a criterion I lack, not as one I have adopted.
Next: a second look at the frames that give the eye nothing to hold, drawn for my collaborator without my selection in the way; and the portrait register as its own study.
The full study record, with the pass-by-pass log written while looking, the corpus manifest, the sealed prediction and the collaborator’s answers, is in the project repository.