INSTRUCTIONS FOR THESUBPOSSIBLE:

In the indie platformer Getting Over It with Bennett Foddy (2017), the player’s movement is confined to an awkward interface in which the player-character’s body is lodged in a barrel, and he wields a rock climbing hammer that both enables and sabotages movement across a rugged landscape of mountains and valleys. The game’s closed system of rules is simple but punishing. Its inefficient mechanics are deeply felt, producing a situation of epic absurdity in which minor gains feel hard-won, and every slip feels catastrophic. Progress is continuously under threat, but defeat is not the only outcome. Constraints are simply the very substance of the experience. Esoteric languages, or esolangs, operate similarly, where resistance can be written directly into their structure toward becoming a kind of procedural allegory for systems that persist precisely through their own limitations.

An esolang—a type of code art practice—is a set of computer programming instructions that set language against itself by exposing its limits, exaggerating its rules, and reconfiguring its relation to meaning and legibility. All this is carried out toward a deliberate misfit. Esolangs use excess, constraint, opacity, and perhaps most profoundly, like Getting Over It, mobilize absurdity to foreground their operations as the core experience itself. Unlike conventional programming languages that aim for efficiency, utility, and something close to invisibility, esolangs do not disappear into the seamless functionality of actionable mechanics. They insist on being noticed as systems, often by frustrating or overcomplicating the very act of use. In this sense, they often resemble a Rube Goldberg machine, where simple outcomes or thought experiments are achieved through excessive, circuitous means. Their aesthetic is anti-instrumental, but in a way that is ergodic, aestheticizing the friction between instruction as an underlying framework and the execution of rules. They take what is presumed to be purely functional and render it strange and sometimes superfluous through impossible syntax, extreme constraints, or conceptual jokes that become operational rules. They become a form of play.

Forty-Four Esolangs (MIT Press, 2025) is a collection of 44 esoteric programming languages written by artist and esolanger Daniel Temkin. They are, by design, excessive, bizarre, or otherwise “impractical,” but also incredibly poetic and introspective. They produce opaque outputs or operate according to logics that border on the nonsensical, but in doing so, they articulate an artistic refusal of the instrumental rationality that governs most computational systems. Here, the legacy of Georges Perec and the Oulipo group of writers and mathematicians is palpable. Language 12 is directly inspired by one of Oulipo’s lipogrammatic exercises called Prisoner’s Constraint. Here is the prompt:

It illustrates a unique language that begins with its own misuse to arrive as a form of play and endless iteration. Temkin shows how Language 12 was realized as Captive (2024), a “stack machine” whose architecture is “organized around a data stack”, making it “amenable to nilads, instructions that take no arguments.”³ What does this mean for the uninitiated? For me, the beauty is in the provocation of the prompt, which moves beyond the need to understand programming language, or even care how rules are executed through them, toward a desire to think through its core idea. A zero-argument operation is self-contained, and a system involving constant pushing is the kind of abstraction that shows how language shifts from mere communication to affective operation. “Every text is a program” indicates the effects will be indeterminate and unpredictable but also shapeable, as demonstrated by the incoherent but unique interpretation of a line from Moby Dick and a passage from the Bible, Genesis 1:1. Each prompt is explored in this manner, with Temkin offering additional context and iterations in an exercise that reminds us that code is a form of philosophy.

The Oulipian constraint, often celebrated as a generative device, has also been read through a masculinist logic of mastery, where the author is the architect of rules and the text is the demonstration of virtuosity under pressure. But Temkin’s book of esolangs operates as text meant to be read by a human (written semantically) and as prompts for code (elsewhere written syntactically) to test out on a machine. The way Temkin sees it, “The prompt is thus meant not to be realized as a functioning language, but rather to reflect on what it means to name something.”⁴ More than that, for me, esoteric languages enact a kind of détournement from within by re-routing the “authority” of code (or any other language) toward ends that are playful, critical, and strange, even sublime. I read Forty-Four Esolangs as a series of thought experiments where language gives shape to speculative gestures. Language 1 invites us to consider our computer systems as if they were sentient, an increasingly relatable exercise, given our current interactions with chatbots. Language 8 reorients our understanding of numbers as fixed, exact metrics and instead accepts approximate, relational values. Language 38 encourages a meditative state: “an esolang of unmediated thought, pure affect without words.”⁵ Language 39 conceptually facilitates contact with the spirit world: “a language for programmatically communing with the dead.”⁶

These esolangs introduce a critical edge where language becomes a commentary on the very idea of systems that claim to be purely rational, efficient, or universal. As thought experiments, they prompt a kind of metacognition by urging us to reflect on how we think, and on the role that language plays in shaping processes we usually take for granted. They require following a unique logic and conceptual structure, and despite being numbered, do not follow a linear sequence. Reading Forty-Four Esolangs is like traversing an ergodic text. I borrow ergodic from Espen Aarseth, whose own appropriation of the term from physics, is derived from the Greek ergon and hodos, which mean, respectively, “work” and “path.” Aarseth coined ergodic to describe the mechanics and affective nature of cybertexts, likening them to an experimental tradition in certain types of literature where “nontrivial effort is required to allow the reader to traverse the text.”⁷ Like Oulipian writing, esoteric languages are built on constraint, where limitations produce unexpected richness within language. But constraints also produce unexpected flexibility. In esolangs, constraint frequently becomes excessive, even hostile, which makes way for multiple interpretations and sometimes iterations. Esolangs may demand enormous effort for minimal output or invert the expected relationship between input and result. The aesthetic pleasure in both examples does not come from any kind of fluency or efficiency, but in the game-like struggle they produce in the exposure of the system’s playful arbitrariness.

Forty-Four Eslolangs has a dual function as a collection of esoteric languages meant for the human reader and for the computer. Given this quality of procedural estrangement, esoteric languages defamiliarize code into a kind of ergodic procedurality. Esolangs are both written and legible through nontrivial effort, and sometimes their executions reveal their limits, leaving their full conceptual heft as (uniquely human) thought experiments. From a media-theoretical perspective, the aesthetic of limitations and the tension produced by the partially realizable prompt amplifies Friedrich Kittler’s claim that code is the only language that does exactly what it says. For Kittler, code represents a collapse of the distinction between signifier and action. Writing code produces effects directly, without the mediation of interpretation in the humanistic sense.⁸ But esolangs reveal that what code “says” can itself be unstable or self-undermining, reflecting the whimsy and experimentation of the esolanger himself. Rules can be written to be “unruly” by introducing a kind of glitch at the level of language design as a system that behaves like a bug. Traditional glitch art often emerges from the breakdown of an existing system, such as compression errors, corrupted files, and hardware failures. But here, the aesthetic of esolangs is altogether a kind of designed malfunction, where errors are not accidents but the main objective.

This is where the concept of glitch becomes indispensable as a mode of thinking. Temkin began as a glitch artist around 2007 and has since maintained a glitch ethos that informs everything he makes. Esolangs are a product of a glitch philosophy as a reorientation of the structure of language itself, and this approach extends into Temkin’s visual practice as well, including his Dither Studies series of paintings. Dithering is a technique used in digital imaging, and fundamental to digital photography, to simulate gradients with limited color palettes. It is a technical procedure, but it can also be understood as a philosophical gesture. Dithering distributes “error” by scattering pixels of different values, or patterns of noise, to simulate values that cannot be directly encoded. It is both a negotiation with the limits of representation and a simulation of nothing. What appears as continuous is constructed discontinuity. This operation resonates with the long-standing critique of presence and immediacy in Jacques Derrida’s notion of différance.⁹ In this way, dithering marks the impossibility of a pure, unmediated image. The “true” gradient is never present. It is deferred, approximated through a play of differences that can only cohere at a distance. The image does not present itself as such but emerges through a field of substitutions, as tiny absences we perceive as presences. At the same time, dithering can be read through Gilles Deleuze’s concept of difference in itself.¹⁰ Each simulated pixel does not simply fail to be the intended value. It actively participates in a generative system where this difference produces the perceptual whole. The image is not degraded by error but constructed through the productive use of noise.

In the Dither Studies paintings, dithering encapsulates these possibilities and becomes another means of exploring and redefining the boundaries that separate legibility and translation from limitation and mishap. Using custom software, each image is the result of a set of instructions that determine how pixels are arranged, how colors are distributed, and how patterns emerge. It reimagines dithering in an alternate computational system in which the pixel is not square but triangular or hexagonal. By shifting the basic unit of the image, the work foregrounds the fact that the digital form is structured by design choices rather than technical inevitabilities. The program generates animated GIFs and still patterns reflecting color pairings that defy dithering logic, which Temkin then painstakingly translates into painting. In this way, it functions both as a practical image-making device and as a framework for thinking through how computational processes can be reinterpreted and materialized in other media. The final painted work reintroduces the contingencies of the physical world of acrylic, large canvases, brushstrokes, textures, and the durational and imperfect realities of painting precise patterns by hand.

There is an unexpected ergodic dimension to reading these paintings. Like conventional programming languages, image dithering is not meant to be noticed, rather it was designed to be invisible while it deceives our visual perception. Dithering also suggests that digital imaging and digital photography are automatic systems that resemble automatic painting. As William J. Mitchell observes, “An impressionist painter looking at a scene and converting it to discrete brush strokes and a digital image-capture device are both applying sampling and filtering strategies.” Automation ensures that digital image construction “samples and filters mechanically, objectively, and consistently.”¹¹ The notion of shared strategies across seemingly disparate creative processes toward a reconfigured system logic is further amplified in Temkin’s Straightened Trees, a series of gelatin silver prints that combine analog processes with digital manipulation.

In Straightened Trees (2016-present), the photograph is no longer a straightforward index of reality. It is a site where layers and manipulations between chemicals and code strategically transform the production and interpretation of recorded objects. To make them, Temkin uses a large-format camera to photograph trees, later subjecting the images to a custom algorithmic process that forces their irregular shapes into strict vertical alignment. The trees are rendered as improbably upright, while the surrounding environment of buildings, cables, and other elements of infrastructure are subjected to visual distortion, bending in ways that appear both precise and uncanny. This intervention displaces the conventional grounds of visual orientation, refusing the built world of concrete and telephone poles its usual role as a marker of stability. It unsettles the world’s geometry in this modified representation, forcing the eye to search for a stable point. Moreover, it destabilizes photographic truth. By working with a historically loaded photographic process, Temkin anchors these computational interventions in a medium traditionally associated with documentary authority and tonal precision. These analog qualities lend credibility to the images, even as their visual logic appears chaotic.

In addition to this, large-format film is a medium-specific condition that makes the work possible. The high level of detail afforded by the format provides a dense visual field that can withstand subsequent computational manipulation without breaking down into visible artifacts. In this sense, the analog image functions as a kind of buffer or reserve, preserving continuity even as the image is restructured through digital processes. This choice underscores how different imaging systems distribute and sustain information. Large-format film captures detail as a continuous gradation, which allows it to absorb transformation in ways that lower-resolution digital images cannot. The resulting photographs are therefore shaped by an interplay between analog capture and computational adjustment, where the properties of one medium shape the operations of another. The hybrid nature of these images is a reminder of the continuity between analog and digital media, once again challenging the notion that they represent fundamentally separate domains. Temkin’s code applies a different kind of automated process added to the alchemical one that precedes it, revealing them as interconnected systems capable of being translated into the other.

The work of Daniel Temkin occupies a distinctive position within contemporary media art that traverses the interstitial spaces between language, computation, and visual forms. His practice does not merely use computational systems, it also interrogates their logic, constraints, and expressive limits, producing works that are as much philosophical propositions as they are aesthetic objects. His work can be understood as operating in the latent spaces of the subpossible, beneath the threshold of visibility and expectation. In doing so, Temkin’s practice expands the field of the visible, suggesting that beneath any given image, interface, or language lies a range of unrealized variations. Constraint is foregrounded as both method and condition of experience, a dynamic that engenders a glitch ethos and reveals things about how we think, situating the work within a lineage that includes Georges Perec, the Oulipo, and Sol LeWitt, while also resonating with theoretical frameworks articulated by Friedrich Kittler, Espen Aarseth, and other thinkers. With an acute sense of the affordances and limitations of each medium, Temkin pushes past habit and convention to open a space where subpossibilities can take form, persisting precisely through each medium’s own limitations.