Vera Molnár programmed uncertainty, then chose by eye

Vera Molnár in her home and studio on rue de Gergovie, Paris, in 1954. François Molnár's medium-format photograph places the artist beneath her geometric collage *Douze Rectangles*, years before she gained access to a computer. Scanned negative, Vera Molnar Archives.[4]

Image context: this is a scan of François Molnár's 1954 medium-format photograph, not a reconstruction or a generated portrait. Vera Molnár is seated in the Paris home that doubled as her studio; geometric works occupy the wall behind her, including Douze Rectangles above her head. The photograph matters because it catches the visual system before the computer entered it: rectangles, intervals, repetition, rotation, and the desk where alternatives could be compared.[4]

The most important machine in Vera Molnár's career did not exist.

Before she gained access to a real computer in 1968, Molnár began working with what she called a machine imaginaire. She would define a limited visual vocabulary, establish rules for changing it, and carry out the permutations by hand. The phrase can sound like a charming prophecy of digital art. It was more rigorous than that. Imagining a machine let her separate making a picture from testing a proposition: what changes when one rectangle shifts, one line tilts, one interval widens, or a little disorder enters a stable field?[1][4]

That distinction is the key to her computer work. The computer did not arrive as an artificial artist that relieved Molnár of decisions. It made variations faster, stranger, and easier to compare. She still decided which rule to write, which parameter to disturb, which result deserved another look, which output should reach paper, and which sequence viewers would eventually see. Molnár programmed uncertainty, then chose by eye.

The studio preceded the computer

The 1954 photograph resists the familiar origin story in which a new device suddenly invents a new kind of artist. Molnár had trained in Budapest, moved to Paris in 1947, and committed herself to geometric abstraction long before a plotter drew for her. Centre Pompidou traces her early work from reduced geometric landscapes in 1946 through postwar Constructivist experiments and randomly distributed elements in 1959.[1]

In the photograph, technology is modest: paper, paint, a desk lamp, a large window, movable shapes. Yet the wall already behaves like an interface. Works can be placed beside one another. Rectangles repeat at different scales and orientations. A composition can be altered without pretending that every alteration is an entirely new inspiration.

Molnár and her husband François sometimes planned paintings and collages together by pinning cut paper to their studio wall and taking turns modifying the arrangement until they agreed it was finished. Zsofi Valyi-Nagy's archival research identifies the large work above Molnár in the photograph as Douze Rectangles, a solo collage now held by the Musée de Grenoble.[4] The important point is not whether a particular rectangle was moved by hand or later by code. It is that comparison, revision, and selection were already physical habits.

The machine imaginaire formalized those habits. Molnár could specify a procedure without owning hardware: begin with elementary forms, enumerate possibilities, vary one condition, inspect the result, continue. This was not a rehearsal for surrendering judgment. It was a way to make judgment encounter more than one answer.

A plotter made the rule leave a mark

In 1968, Molnár began using a computer at a Paris research laboratory and learned FORTRAN. The machines were institutional equipment, not studio appliances. Artists seeking access to computing in the 1960s commonly depended on scientific laboratories and technical collaborators; the V&A places Molnár among the pioneers who converted that restricted access into a new form of drawing.[2]

Her series Interruptions shows what the transfer from rule to paper could produce. The V&A's 1969 sheet is filled with short black lines whose directions vary. Across that dense activity, irregular white openings appear. They are not illustrated objects. They are absences generated inside a field: places where the procedure has been interrupted.[3]

The sheet complicates the usual opposition between mechanical precision and expressive accident. Each short mark is clean enough to disclose the plotter's controlled movement, but the whole refuses the visual comfort of a regular grid. The gaps do not resemble one another closely enough to become a pattern. Order supplies the pressure; local deviation makes the surface breathe.

The V&A describes Interruptions as an iterative series: Molnár slightly modified the algorithm from one work to the next, using each result to alter the next possibility.[3] That sequence matters more than the romance of a machine producing a surprising picture on command. One output was evidence for another trial. The artwork developed through a loop.

The screen shortened the distance between question and answer

Early computation could make that loop painfully slow. Molnár recalled sending instructions without seeing an image and receiving a result a day or several days later. A bad assumption survived until the batch returned. When she encountered an IBM 2250 graphics display, the screen changed the tempo: parameters could be adjusted while an image was visible rather than after paper output had arrived.[4]

Her 1975 paper “Toward Aesthetic Guidelines for Paintings with the Aid of a Computer” describes the method with unusual clarity. Simple forms were altered successively on a cathode-ray-tube display. When an image seemed worth preserving or translating into a painting, she could have it recorded by a plotter. By changing a series in small steps, she tried to identify which alteration made one result feel stronger than the preceding ones.[5]

This was computation used as an instrument of discrimination. The screen did not provide a final picture so much as a fast provisional one. It allowed Molnár to ask another question before attention cooled. The plotter then changed the encounter again. Its pen moved through real time across paper, leaving ink; many surviving sheets also retain sprocket holes, margins, timestamps, or Benson markings. Image, instruction, and object were connected, but they were not interchangeable.[2][4]

Molnár called this a “conversational” method.[4] The metaphor should not be inflated: the computer did not understand her, and its outputs were bounded by her program and the available machinery. But conversation is accurate at the level of revision. She did not write a rule once and accept all its consequences. A result could cause her to rewrite the rule.

Randomness still had an editor

“Generative” can imply that the originating procedure possesses the work's authority. Molnár's practice makes that description incomplete. She used randomization and controlled percentages of disorder, but she did not treat every permitted output as equally successful. Centre Pompidou's survey of her 1970s work emphasizes the repeated introduction of disorder into simple geometric compositions; her own 1975 account makes aesthetic comparison the purpose of generating the series.[1][5]

The 1974 plotter suite Hommage à Barbaud exposes the editorial layer. Its concentric squares move from a disciplined grid toward displaced centers, altered sides, and apparent tangles. Valyi-Nagy reconstructed the production sequence from marginal machine text and sprocket-edged sheets. The order in which Molnár exhibited the selected images did not simply reproduce the order in which the computer made them. She chose a small group from many variants and arranged that group into a persuasive progression.[4]

The disorder, then, operated at several levels. Random values disturbed individual forms. Molnár disturbed the program as she learned from its results. Finally, selection and sequencing disturbed chronology. What looks like an algorithm unfolding inevitably is partly a montage assembled after the computation.

That does not weaken the work's systematic character. It identifies where authorship lives. A procedure can open a field of possibilities without deciding how that field should be experienced. Molnár's eye supplied thresholds the code did not contain: enough tension, too much regularity, an unexpectedly alive interval, a variation worth keeping beside another.

The paper remembers what museums once overlooked

At the V&A, computer art entered through categories built for older objects. The museum notes that early digital works were often collected by prints-and-drawings departments according to their paper support; some plotter originals were represented by screenprints. When the museum acquired three Molnár works in 2011, calling them “plotter drawings” shifted attention from appearance alone toward the process that produced the marks.[6]

That change in naming repairs only part of the history. A finished sheet cannot preserve the program, discarded outputs, screen interactions, laboratory access, or the artist's changing criteria. Conversely, preserving code without the paper would erase the line, scale, surface, and sequence through which the work entered a room. Molnár's art belongs to neither software nor drawing alone. It occupies the relay among rule, display, mechanism, hand, eye, and archive.

Her career also outgrew the narrow label “computer artist.” Centre Pompidou's 2024 retrospective included notebooks, photography, painting, plotter drawings, sculpture, and large installations, following recurrent lines and geometric families into work made as late as 2023.[1] The consistency was not a single device. It was a way of returning to forms and making them yield another difference.

This is why the 1954 studio photograph is more than prehistory. The rectangles on the wall do not await the computer to become intelligent. They show that Molnár had already made the studio into a place where a form could be varied, compared, rejected, and resumed. Hardware later expanded the speed and range of that inquiry. It did not eliminate its pauses.

Molnár's enduring lesson for generative art is therefore neither that machines are creative nor that only human taste matters. It is that authorship can be distributed across distinct acts without becoming vague. A rule initiates. A machine calculates and draws. An output resists expectation. An artist notices, alters, keeps, and orders. The work becomes visible in the passage between those verbs.

cronfeed.work