Derek Jarman’s “Blue” (1993) is a movie without an image. For 79 minutes, it shows only one frame, filled entirely with dark blue, devoid of detail. And yet it is a profoundly cinematic experience: the world unfolds before the spectator, even though it is invisible.

Nothing can be seen in “Blue”. The only way to navigate it is by following the ambient sounds and the narrator’s voice. Yet the world gazes back, acting upon you. This is what images do: they reach us, impress us and eventually set our bodies into motion.

This essay is about the physical power of images, a force inherent to them even when they are not visible.

Paradoxically, I suggest that the visual aspect of an image is not intrinsic to its nature, but rather a possible byproduct of its production. Images are not born to be visible at all. Some are — as in photographs, paintings, or digital interfaces. Yet others remain invisible to the human eye, without being any less images.

I take my cue from the philosopher Peter Szendy, who sees images as archaeological configurations: diverse temporal fluxes held static¹. An image is a living palimpsest. It is structured by time, frozen forever, and yet somehow dynamic.

Time is only perceivable when it applies to something. An apple falls and hits a man under the tree; a child’s face grows longer and the first pimples appear; a dictator dies, and the revolution follows. These are the agents and the events affected by time. The reverse also holds true: time manifests itself through them. Bodies set time in motion, as much as the time moves the bodies.

The temporal fluxes that constitute an image are also fluxes of power and capital. Images feed on living bodies. They have always watched us, shaping us through their gaze, and there is no refuge inside the camera obscura. There is no observing without being observed: the disembodied Cartesian mode of seeing is a great ideological myth of modernity.²

1. What Makes a Photo a Photo?

A photograph is always invisible, wrote Roland Barthes in “Camera Lucida”³. What he meant was that photography’s referential power conceals its materiality. Faced with a portrait, we first see a face — as if a living person stood before us. And only later, after this stage of recognition, do we notice the photographic properties of an image: its composition, the grain on the film, a vignette.

“Camera Lucida” was Barthes's final book: he died two months after its publication in 1980, after being struck by a car in Paris. It is unclear whether he sensed the shift, but in his final years some images became invisible in a literal sense — not merely metaphorically, as in his provocative claim. It was the moment when digital images started to proliferate, slowly supplanting the analogue ones.

In 1977, the Atari 2600 was released — the first truly popular home videogame console. Plugged into the TV, it generated playable images. These arose from a realm untouched by sunlight: the binary operations of the console’s motherboard. To be seen, they had to be externalised by the Television Interface Adaptor — and drawn line by line by an electron beam.⁴

As the media theorist Seth Giddings puts it, digital images are drawn without light.⁵ This contrasts with photography, which is, according to its inventor Henry Fox Talbot, an imprint left by the sun for us to observe. Even if digital images are visible, they do not necessarily belong to the visual. They just happen to be externalised in a visual modality, but are in no way bound to it.

By the end of the 20th century, computational capacities had grown exponentially, enabling consoles to render three-dimensional images in real time. Digital “cameras” had become full-fledged characters — as in first-person shooters. And even though, technically speaking, they had nothing in common with optical cameras, they ended up thoroughly imitating an analogue lens.

The first fully 3D shooter, Quake (ID Software, 1996), immersed players in a world of polygons — each with a position in three-dimensional space, a texture, and lighting properties. The game’s engine calculated perspective according to Renaissance rules for multiple objects, repeating the process dozens of times per second. Its successor, Quake III Arena (ID Software, 1999), also simulated the optical effects of real cameras: bright light sources produced lens flares. More recent games added motion blur, chromatic aberration, and lens distortion to the list.

One could argue that videogames created an operational remediation of the camera obscura, which is essentially an observation technology. It allows one to see space on a plane — to perceive the order of things before the viewer without being a part of it. Through billions of operations, a digital camera obscura produces a similar spatial illusion on the screen.

One imitates only that which one is not. Consequently, a digital photo is not a photo. Digital imagery comes from a non-visual operational realm, while photography seems to be a primordially visual medium. And yet I suggest that, at a deeper level, digital images are still images.

A screenshot — just like a photograph — has indexical power, as Giddings points out.⁶ It is evidence of an event that happened. An optical image is made of living (and moving) bodies: a flux held static. A digital one is made of operations — which, like bodies, unfold in time, interact with one another, and form multiplanar causal chains.

Even if a digital image is not visual, it is still a palimpsest: composed by multiple layers of frozen forces, static yet struggling to break free.

2. What Makes a Map a Map?

In 1985, Vilém Flusser introduced the concept of the technical image.⁷ Such images fuel cybernetic machines: computers think with them, just as we think with thoughts. Made of codes, they stand in opposition to alphanumeric writing, a cultural technique that, according to Flusser, gave rise to history in the human sense.

Writing was one-dimensional: a linear sequence of phonemes, where one precedes the next. Traditional pictures, as Flusser calls them, show an object on a two-dimensional surface. Technical images, by contrast, lack dimensionality altogether. They are medium-free assemblages of information, capable of incarnating in any modality — visual, textual, or sonic.⁸

At first glance, Flusser’s concept appears to have much in common with Harun Farocki’s notion of operational images. The latter, according to the artist, are not made simply to represent things but rather to operate on the material world through techno-social infrastructures. They are part of larger operations — and, first and foremost, military ones, such as images from guided bombs or footage used in facial recognition systems.

But are there any non-operational images? Flusser suggests that representation is not the ultimate goal of images; we represent things in order to interact with them. Images grasp the world because the eye grasps more than the hand can. They are flat because the eye scans surfaces, allowing us to grasp more.

Purely representational images are a myth of modernity — born inside the camera obscura. Images are not passive, and never were; they are made to act. Consequently, every image is operational.

The difference is that some images represent things and bodies. A hand-drawn portrait, a photograph, an emoji: all of them depict some easily perceivable element of the visual world. Other images, though visible to the eye, do not refer to any visual entity. These are maps and diagrams: they are made not to be contemplated but to be followed.⁹

Hence, non-visual images predate cybernetic technical images. A grid — with the map as its special case — is made not to be seen but to serve as a direct instruction for action. Furthermore, as a cultural technique, it was meant to serve not individuals but a broader and more abstract infrastructure built from human bodies: the state. To paraphrase James C. Scott, the state sees the world in grids: in the spreadsheets that track the citizens’ spatial and legal status.¹⁰

I would suggest that maps and diagrams were the true precursors of technical images. Here I follow Matteo Pasquinelli, specifically his idea of the social genesis of AI algorithms.¹¹ He argues that cybernetic machines mimic the social principles of the division of labor, which go back to the first factories.

In a similar fashion, the state feeds on non-visual images — grids, maps, and diagrams. Their power organizes human bodies in space and time, just like technical images, which orchestrate operations within computers.

3. What Makes an Image an Image?

Today’s networked image is a case study of Flusser’s technical image. Cybernetic machines mediate quasi-seamlessly with one another and with users’ nervous systems. Together they form the net, where space is subsumed by time. Once an event reaches it, it spreads to all its nodes, no matter how distant.

To contemplate a networked image means to be contemplated. Once a user encounters a picture on social media, there is no doubt they are observed by the cold gaze of AI algorithms. These are not visible, nor do they see in the human sense, nor do they need sunlight to navigate, as we do. And yet they observe, softly exercising their power on us — a power that is ultimately corporate or state, filtered through a statistical lens.

Traditional images were made to grasp the material world. As Flusser suggests, they cast a spell on objects to keep them at human disposal. It turns out, however, that the spell had a flip side: it equally allowed the world to grasp us. The networked image is a perfect illustration of this logic. Yet the same holds true for any other image.

Images grasp and trouble us, even provoking bodily effects. Consider two of cinema’s most “bodily” genres: horror and porn. They crawl under our skin, giving us shivers of pleasure or disgust, sexual excitement or the urge to vomit.¹² This is how the world affects us through an image: never passive, and never made for contemplation alone.

An image is made of bodies in motion — whether a picture, a static photograph, or a movie. It condenses the heterogeneous flows of time into a flat surface. And yet it is never a still life, because life fails to be still. Rather, it is a transient sum of the vital powers¹³ of the material world: static, and yet dynamic in its stasis.

For this exact reason, images are not inherently visual: they are defined by a paradoxical configuration of forces. A flat, two-dimensional surface happens to be a perfect place for the resolution of these momenta. But some images never arrive there, lingering in the dark — as many technical images do.

Peter Szendy advances the idea of an ecology of images, driven by an urge to account for their vital powers. For him, images are akin to living organisms, which means they never exist on their own but combine into larger ecosystems.¹⁴ To unfold such an ecology is also to push the borders of visual culture beyond the Anthropocene: to see beyond human and human-based temporalities.

I propose a simpler idea. Before conceiving images as organisms, we should understand their behavior as bodies. I’m advocating for a Newtonian physics of images: a tentative foundation for a new critique of visual culture, one that accounts precisely for the physical power of images.

An image retains the momentum of its conception. Once seen, that momentum reaches another body — that of the observer. According to the fundamental laws of classical mechanics, energy is conserved over time. Thus, the observer’s body cannot remain intact.

When seen by the eye, the momentum of an image is resolved on the skin. We feel it as a shock, or a gentle touch, or a chill, or a spasm.

This is yet another way images have always been invisible.

A Szendy, P. (2025). For an Ecology of Images. Verso Books.
cf.: Crary, J. (1992). Techniques of the observer: On vision and modernity in the nineteenth century. MIT press.
Barthes, R. (1980). La chambre claire. Gallimard.
Montfort, N., & Bogost, I. (2009). Racing the beam: The Atari video computer system. MIT Press.
Giddings, S. (2013). Drawing without light: Simulated photography in videogames. In M. Lister (Ed.), The photographic image in digital culture (2nd ed., pp. 15–30). Routledge.
ibid.
Flusser, V. (2011). Into the universe of technical images (N. A. Roth, Trans.). University of Minnesota Press.
See also: Flusser, V. (2011). Does writing have a future? (N. A. Roth, Trans.). University of Minnesota Press.
Siegert, B. (2011). The map is the territory. Radical Philosophy, (169), 13–16. See also the review in: Parikka, J. (2023). Operational images: From the visual to the invisual. University of Minnesota Press.
Scott, J. C. (2020). Seeing like a state: How certain schemes to improve the human condition have failed. Yale University Press.
Pasquinelli, M. (2023). The eye of the master: A social history of artificial intelligence. Verso Books.
Marks, L. U. (2000). The skin of the film: Intercultural cinema, embodiment, and the senses. Duke University Press.
Szendy, P. (2025). The ecology of images: An aborted idea? In For an ecology of images. Verso Books.
Compare with an account of the ecologies of platform images in: MacKenzie, A., & Munster, A. (2019). Platform seeing: Image ensembles and their invisualities. Theory, Culture & Society, 36(5), 3-22.

References

Crary, J. (1992). Techniques of the observer: On vision and modernity in the nineteenth century. MIT Press.

Flusser, V. (2011). Does writing have a future? (N. A. Roth, Trans.). University of Minnesota Press.

Flusser, V. (2011). Into the universe of technical images (N. A. Roth, Trans.). University of Minnesota Press.

Giddings, S. (2013). Drawing without light: Simulated photography in videogames. In M. Lister (Ed.), The photographic image in digital culture (2nd ed., pp. 15–30). Routledge.

MacKenzie, A., & Munster, A. (2019). Platform seeing: Image ensembles and their invisualities. Theory, Culture & Society, 36(5), 3–22. https://doi.org/10.1177/0263276419847508

Marks, L. U. (2000). The skin of the film: Intercultural cinema, embodiment, and the senses. Duke University Press.

Montfort, N., & Bogost, I. (2009). Racing the beam: The Atari video computer system. MIT Press.

Parikka, J. (2023). Operational images: From the visual to the invisual. University of Minnesota Press.

Pasquinelli, M. (2023). The eye of the master: A social history of artificial intelligence. Verso Books.

Scott, J. C. (2020). Seeing like a state: How certain schemes to improve the human condition have failed. Yale University Press.

Siegert, B. (2011). The map is the territory. Radical Philosophy, (169), 13–16.

Szendy, P. (2025). For an ecology of images. Verso Books.

artworkfieldwareIan Margo 2025, digital video, color, sound 16:9 4K, 11:00min...(more)