“Epistemological Machines” | @ MGLC Švicarija, Ljubljana
By Lea Sande & Tisa Troha
Transcription
Maks: I'm really glad to introduce Tisa Troha and Lea Sande. Tisa is an architect, editor of Šum Journal, the girl who introduced girl-theory to Ljubljana, and perhaps the original Hegelian e-girl. Because Claire was so good with introductions yesterday, I have to level up my game. Lea Sande is the most prominent new voice of the new generation of thinkers and artists in Ljubljana. She is an editor of the Multithread series by Aksioma, host of the Feedback Loops show at Radio Študent, and is a theorist and writer herself.
Tisa: When we hear the term black hole, it's easy to imagine a mysterious cosmic object, something out there in space that we might, nevertheless, be able to see more clearly one day when we have the technology for it. But there is a catch. Black holes are, by their very nature, not directly observable in themselves. We arrive at the fundamental contradiction, where we can never actually see a black hole itself, and yet, counterintuitively, they're among the most carefully studied and mathematically modeled objects in contemporary physics. So how is that possible? What we know about black holes is not derived from direct perception, but from the interpretation of their indirect effects — radiation, motion, and gravitational distortion occurring at their boundary. We can discuss, calculate and simulate the singularity — the paradoxical point of infinite density at their core, where spacetime curvature becomes infinite — and its surrounding event horizon, the boundary beyond which the escape velocity is greater than light speed, and thus nothing, not even light, can escape. The question we're interested in today is not so much what black holes are, but how they become knowable at all in their impenetrability.
In this talk, we propose a different way of thinking about black holes. Instead of treating them only as astrophysical objects, we theorize them as something we might call epistemological machines — structures that organize the conditions for knowledge production, and more specifically, that produce knowledge through the interpretation of a visible surface, the event horizon, as evidence of, and material for inference about, a fundamentally invisible and inaccessible absence: the singularity. Drawing on philosophical accounts of holes, the concept of the black box, and the machine of faciality, we can generalize the relationship between what we can observe and what we can never directly access to things other than black holes — to a more general pattern in how knowledge works. We propose that the relation between event horizon and singularity exemplifies a general mechanism of knowledge production: a visible surface that makes possible the inference of an inaccessible center.
On holes. This structure becomes clearer when placed in dialogue with philosophical discussions of holes. As shown in the dialogue by David and Stephanie Lewis, the status of holes is deeply paradoxical. On the one hand, we routinely speak of holes as if they exist: they can be counted, compared and located. On the other hand, holes are not material objects; they are defined precisely by the absence of matter. This produces a tension between language and ontology: holes behave like entities, yet seem to lack the substance required to qualify as such. Achille Varzi sharpens this point by observing that a hole is "there where something isn't." Holes are not nothing, but neither are they things in the conventional sense — they are structured absences, entities defined relationally through the configuration of the material that surrounds them. They have shape, location, and persistence, yet they cannot be reduced to the matter that hosts them. In this sense, holes expose a limit within ontology itself: they force us to acknowledge that absence can be differentiated, organized, and operationalized. Attempts to eliminate holes by reducing them to properties of objects — treating "there is a hole" as equivalent to "the object is perforated" — fail in important cases. As Varzi shows, topological descriptions of objects cannot fully account for differences between types of holes without implicitly reintroducing them. The conclusion is then unavoidable: holes cannot be reduced away. They must be taken seriously as part of our ontology. What they reveal is that absence is a structural feature of reality rather than a lack.
We can map this framework directly onto black holes. A black hole is defined by an absence: nothing, not even light or any other carrier of information, can escape from it. Yet it is treated as a real, measurable entity. Like ordinary holes, it is not reducible to the material or spatial structures that surround it. Instead, it emerges as a structured absence that organizes those structures. The event horizon marks the boundary of this absence, allowing it to be located and analyzed without ever being directly accessed. Black holes thus represent a limit-case in which absence becomes the primary condition of knowledge.
Lea: In classical epistemology, knowledge is often understood as grounded in observation. Objects are known because they can be perceived, measured, and described. Black holes disrupt this model. The singularity at their center is, by definition, unobservable, as we've already established. What is accessible to observation is limited to the boundary — what happens around the black hole, not within it. Yet black holes are not treated as speculative or fictional entities; their properties — mass, spin and location — are calculated with remarkable precision. This produces a paradox: how can something fundamentally unobservable be so thoroughly known? The answer lies in the structure of inference. Scientific knowledge of black holes is generated through the interpretation of data collected at the level of the event horizon. The horizon functions as a surface on which observable phenomena are inscribed, allowing scientists to infer the existence and properties of the gravitational singularity. Knowledge is thus not derived from the object itself but from a structured relation between surface and absence.
At this point, the concept of the black box becomes crucial. As René Thom argues, scientific practice is grounded in opaque systems whose internal functioning remains inaccessible. The task of science is therefore to explicate, to unveil these black boxes through interpretation — or, as he puts it, "the only conceivable way to expose a black box is to play with it." Scientific knowledge is fundamentally hermeneutic: it proceeds by interacting with systems whose interiors cannot be directly accessed, and whose structure must be inferred through their effects. Thom adds that science conceived this way may be better positioned to answer Heidegger's cold 1929 judgment that science does not think. He distinguishes between two scientific methods by which this interpretation can proceed: the largely limited, analytic method of breaking open the black box and peering inside it, and the hermeneutic method of almost game-like interactions and identifications with the spirit in the box — in Thom's own words, "the ghost in the machine."
This epistemological distinction is re-inscribed and further refined in Alexander R. Galloway's distinction between two other types of black boxes. The first is the black box as a cipher, which can be opened, decoded, and ultimately understood by revealing its internal mechanics. The second is the black box as a function, which remains irreducibly opaque while still being operationally usable — the system provides just enough transparency to function, but never enough to fully disclose its inner structure. Galloway's reformulation of Marx's "rational kernel inside a mystical shell" into "a rational shell and a mystical kernel" marks a shift from a model of eventual demystification to one of permanent, structured opacity — a black box as a mythical ghost in a rational machine. Black holes align decisively with this second model, because they are black boxes that cannot be opened and fully understood, since they must remain opaque in order to be known at all. Their interior is fundamentally inaccessible, and yet their behavior can be modeled with increasing precision.
This perspectival structure has deeper implications. In the case of black holes, black-boxing appears to loop back on itself, bringing into relation theoretical frameworks that would otherwise remain incompatible. For example, tensions between quantum gravity and conformal field theory are not fully resolved, but are made to co-exist through shared formal structures, such as the correspondence between conformal field theories and Anti-de Sitter spaces. Black holes thus function as sites where scientific paradigms are not unified but forced into relation, generating knowledge through their interaction rather than their reconciliation. Black holes should therefore be understood as black boxes in the strongest sense: systems that both limit transparency and reorganize the conditions of explanation, actively structuring how knowledge can be produced. As such, they transform epistemology itself, shifting it away from ideals of transparency and toward a model based on opacity, mediation, and probabilistic inference. In this sense, black holes may be understood as the most compelling and complex black boxes of all. As Iris Long writes, we should catch sight of the black box as an unformed spacetime, a field of proto-probabilities.
Tisa: So, at this point we can finally introduce Deleuze and Guattari's concept of the white wall / black hole into the equation. In A Thousand Plateaus, they describe the face as a system composed of two elements: a white wall that serves as a surface of signifiance, and a black hole that performs as a point of subjectification. The face is a machine that organizes meaning by projecting signs onto an exterior surface and assigning them to an interior subject. As they put it: the face constructs the wall the signifier needs in order to bounce off of, constituting the wall of the signifier, the frame or screen; and the face digs the hole that subjectification needs in order to break through, constituting the black hole of subjectivity as consciousness or passion, the camera, the third eye. What we're interested in here is, of course, not the specific example of the face, but the structure it reveals — a visible surface that organizes signs pointing toward an invisible center. As Deleuze and Guattari put it, the black hole/white wall system is, to begin with, not a face but the abstract machine that produces faces according to the changeable combinations of its cogwheels; we shouldn't expect the abstract machine to resemble what it produces or will produce.
This is precisely the structure at work in the epistemology of black holes. The event horizon works as the white wall: a surface on which observable phenomena are inscribed. The singularity occupies the position of the black hole: an invisible center that is never directly accessed but is nonetheless necessary for the interpretation of those phenomena. What follows from this is that the black hole/white wall system, as applied here, can perform as a conceptual schema for how knowledge operates whenever it's confronted with limits — not just in astrophysics. The point is not that there is something hidden behind the surface that we might one day reveal, but that the very distinction between surface and depth is produced by the machine itself. The "interior" is constituted through inference rather than preceding it. In the case of black holes as theorized by special relativity, it can be tempting — but false — to view the singularity as an actually existing object waiting to be discovered, rather than as an inferred necessity that stabilizes the interpretation of what appears at the event horizon. Without the hypothesis of the singularity, the observable phenomena — gravitational lensing, accretion disk radiation, relativistic motion — would not cohere into a consistent explanatory framework. More than just an object of knowledge, the black hole is, in this sense, a condition for the possibility of a certain kind of knowledge that depends on the continuous translation of surface effects into statements about an absent center.
Perhaps we can extend our initial claim at this point: instead of revealing what is hidden, epistemological machines produce a relation between visibility and invisibility as such. The white wall and the black hole emerge together as a coupled structure. The surface organizes and amplifies signs — patterns, signals, perturbations — while the inferred center functions as a point of convergence that gives these signs coherence. Knowledge arises in the relation between the two. Importantly, this relation is never resolved. Just as the singularity remains inaccessible behind the event horizon, the "center" of any epistemological machine remains structurally out of reach. This is a constitutive feature of the system rather than a limitation to be overcome. Attempts to fully "open" the black box would dissolve the very conditions that make it intelligible. If the singularity were directly observable, it would cease to function as a singularity in the epistemological sense and become just another object among others, as predicted and projected by the exotic theories of quantum gravity attempting to bridge general relativity with quantum mechanics, such as the various string theories.
We can now generalize this structure beyond astrophysics. Many other domains of knowledge rely on similar configurations: a field of observable traces that are interpreted as evidence of an underlying, inaccessible process. In psychoanalysis, symptoms are read as surface expressions of an unconscious that cannot itself be directly accessed. In machine learning, outputs are interpreted as the effects of high-dimensional models whose internal representations remain opaque. In each case, a white wall/black hole system organizes the production of knowledge. What marks black holes as epistemological machines is the clarity with which they stage this structure. Nowhere else, it seems, is the impossibility of access so absolute, and the reliance on inference so explicit. The event horizon is a literalization of the epistemic boundary: a limit beyond which observation is not difficult but in principle impossible. At the same time, it is precisely this limit that makes the system so productive — the impossibility of direct knowledge forces a proliferation of indirect methods: mathematical modeling, simulation, analogical reasoning, through which the invisible is continuously reconstructed.
This leads to a final implication. If epistemology has traditionally been oriented toward the ideal of transparency, toward the progressive elimination of opacity, then black holes suggest a different paradigm. Opacity is not an obstacle to knowledge but its enabling condition. The black hole marks the point at which knowledge becomes reflexive about its own limits and procedures. It is here that epistemology becomes machinic: a system that produces reality as knowable through specific operations of mediation, inference, and projection. To think in terms of epistemological machines, then, is to shift focus away from the question "What is there?" toward "How is whatever thing is there made knowable?" Black holes provide a particularly striking answer: what is most real may be precisely what cannot be seen, and what cannot be seen may only exist for us through the structured play of surfaces that both reveal and conceal it.
Lea: The epistemological structure outlined above becomes fully concrete when situated within the framework of Benjamin Bratton's theory of planetary computation. For Bratton, contemporary systems of sensing, modeling, and representation move beyond operation at the scale of individual instruments or observers to the scale of the planet itself. Computation is distributed across infrastructures that integrate satellites, sensors, data centers, and algorithmic processes into a single, operational surface. Lukáš Likavčan's notion of comparative planetology further sharpens this point: he emphasizes that planetary thinking allows differences to coexist within a shared framework of comparison. This is particularly relevant in the case of black holes, where general relativity and quantum mechanics remain fundamentally incompatible.
The first image of a black hole exemplifies this transformation. As Bratton notes, "the planet itself became the camera" — the Event Horizon Telescope, a network of telescopes distributed across the Earth, collectively functioning as a computational apparatus capable of synthesizing data from multiple points into a unified image. What appears as a coherent visual representation is therefore the product of a planetary-scale process of data aggregation, synchronization, and interpretation. Through algorithmic modeling and the use of neural networks, petabytes of data could be synthesized into a single constructed image. The data collected by the Event Horizon Telescope consists of radio signals that bear no inherent visual form; there is, in principle, no part of the process that guarantees the patterns recognized in the data will have recognizable form. The recognizable "ring" of the black hole emerges only through layers of modeling and algorithmic reconstruction. These processes are themselves partially opaque, relying on computational systems whose inner workings are not fully accessible even to their operators. In this case, planetary computation operates through a stack of black boxes, each mediating between raw data and interpretable output.
As Saša Grozdanov and Eszter Polonyi argue, this shift also reconfigures the role of the observer. In classical models of knowledge, the observer occupies a stable position outside the object of study. In planetary computation, this position is distributed and embedded within the system itself. Knowledge of black holes emerges only through the combination of multiple observational viewpoints, coordinated across the globe. No single perspective suffices; the observer takes up a networked position within an infrastructural system. Black holes already function as black boxes in Thom's sense: systems whose interiors cannot be accessed directly and must be interpreted through their effects. Planetary computation adds an additional layer of opacity, as the processes that transform data into knowledge are themselves black-boxed. The result is a recursive structure in which black boxes produce knowledge about a black hole, itself already defined by absence.
Tisa: Black boxes reveal a fundamental feature of knowledge: it is often produced through the interpretation of structured absences, without ever being able to fully peek into the black box. The relation between event horizon and singularity exemplifies a general epistemological pattern in which a visible surface makes possible the inference of an invisible center. By bringing together Deleuze and Guattari's concept of the white wall / black hole, philosophical accounts of holes as structured absences, the mechanics of black boxes, and contemporary analyses of planetary sensing systems, we can understand black holes as epistemological machines in a precise sense. To think in terms of epistemological machines is therefore to abandon the ideal of transparency. What black holes ultimately show is that knowledge progresses by learning how to operate within opacity.
©2026, Berlin/Nicosia
hello@becoming.press
@becoming.press
Prolegomena
Impressum
Submissions
Distribution
Imprints
(un)Becoming
Agenda
Contact
Newsletter
Subscription
Events
Radio • On-Demand
ÊT/RE • Soundcloud
Materials
Immaterials
Seminars
Cinematics
Nodes
Search