Paraliminal Conceptuality and the Abstract of Infinity, or Film Philosophy into LLMs Will Do Fine

This article questions whether there exists a difference between the actuality and the virtuality of land as a means of image or scapeness with regard to Ben Koder’s Looking Glass Quilt and John Power’s work on generative ambient screens in public spaces as encounters. It also challenges, but more along the lines of plays with, Jeff Malpas’s contestation of space as a concept that is central to the notion of geographical thinking in the absence of a geography, but with an emphasis on the relational view of space that has come to dominate geography and the social sciences as an elucidation of space itself. Indeed, the concept of space as place has come to represent the grounds for its own virtualisation in media and communications research, and to a larger extent through the recent digitisation and manipulation of the image via drone, rather than the perception of that image as an index, or the experience of that image as a spatialisation—a paraliminal conceptuality on the precipice of film philosophy, and a new, interdisciplinary field of research growing in response to the increasingly qualitative aspects of the computer sciences as they converge on topics of Language Learning Models (LLMs) and Artificial General Intelligence (AGI). Tasked by the Virtual Experiences Laboratory (VXLab) at RMIT University, Ben Koder’s most recent experiments with mixed realities, machine learning, and games design, has led him to a qualitative crossroads to do with photography, image manipulation, and what the late Tom Gunning once referred to as “the truth claim of traditional photography (and to some extent cinematography) which has become identified with Charles Peirce’s term ‘indexicality’” (39). Gunning also makes the claim that, at least at the level of the concept of the image, the point of an index makes little to no difference between celluloid and the digitisation “of a physical relation between the object photographed and the image finally created” (40). The one is merely a chemical reaction to, the other a numerical representation of, the same datapoint—the object being photographed—but that celluloid can better assume the difference between the actual and the virtual, object for objet d’art, including the physical relation between the object being photographed with the (re)production of another object. Koder finds the prospect of such a distinction frustrating, for the STEM College at RMIT tends to encourage a much more quantitative approach towards problem-solving that has ultimately left him pursuing an Honours in Media and Communication on my advice as a fellow in the Lab. It is this lasting, quantitative approach that current film studies tend to exacerbate whenever cinematography becomes too (film) philosophical, too, the so-called ‘indexicality’ of claims being made by the one exceeding the scope of the other, and, at the risk of my straying from the topic of land, landings, landscape altogether, the less than qualitative reach of an uncertain grasp that the computer sciences struggle to anticipate about LLMs and the question of consciousness. That not everything real can be quantified, yes, but that not everything quantifiable can be realised in any actual way beyond the virtual, or the philosophical. This newfoundland, as it were, begs the question of a paraliminal conceptuality in which Koder’s experiments verge on the tip of a new, interdisciplinary iceberg. Gunning might have us put aside the concept of ‘indexicality’ for just a moment, in this case the use of a drone, rather than the camera on the drone, which enables the object to be photographed, despite that “what passes for progress (especially theoretical progress) often simply displaces unresolved problems onto new material” (39). In this sense, there is nothing all that dissimilar to a drone than to a change in perspective, one that could not also be achieved with a crane, or a plane, or a helicopter; although Gunning would probably draw the line at something like a John Olsen painting of Lake Eyre or a Mavis Ngallametta painting of Ikalath because the one is only an impression of the other. Indeed, the transfiguration from eye in the sky to paint in the brush is fraught with its own phenomenology of perceptions that, unlike cinematography, say, and Deleuze with the Cinema books, remains contested at the level of the concept of the image (notwithstanding Barthes or Sontag; or, as in the case of Ngallametta and Indigenous Australian cartographic paintings, whether or not the impression of a bird’s eye view is constitutive of an actual bird’s eye view). Vermeer almost precludes this debate entirely, and despite his purported use of optics while painting as per Teller’s Tim’s Vermeer (2013). “But what problem does this change present”, as Gunning himself puts it, “and how does it change indexicality?” (40). Consider not the perspective, then, but the ease with which a drone allows for such a point of view, and that, much like Deleuze’s interpretation of a bourgeoise woman who starts to “see” the world around her in Rossellini’s Europe ’51 (1962), again those “situations which we no longer know how to react to, in spaces which we no longer know how to describe” (xi). We are sat in front of Building 4, Bowen Street, for example, formerly the old Trade School building before The Working Men’s College became The Royal Melbourne Institute of Technology (RMIT). There the drone sits. Not as clumsy or random as a Gaspard-Félix Tournachon taking pictures of Brussels while hanging out of a hot air balloon (as caricatured by lithographer Honoré Daumier), truly a precision instrument for a more binary age. In a little under 15 minutes, though, Koder will have digitised more than just a series of snapshots of the Melbourne CBD (see fig. 1). He will have produced a 3D Gaussian Splat (3DGS) soon to be reminiscent of Tournachon’s own Revolving self-portrait. Fig. 1: April 2024, the original perspective as seen by the drone at 0 degrees, what Koder refers to as the 'Ground Truth' It is the temporal aspects to this Splat, in particular, that intrigue me the most as a film philosopher, what with my own thesis on time, affect, and the moving image. That time can (and does) express itself as itself spatially, and yet, invisibly, inimitably, inseparably, if only briefly, as a time and a space through us—through the affection of the image—a spiel about Deleuze that Koder is more than familiar with my talking about. Deleuze would doubtless attribute this affection to the “loosening of the sensory-motor linkage, [that] it is time, ‘a little time in the pure state’, which rises to the surface of the screen. Time ceases to be derived from the movement … . Even the body is no longer exactly what moves; subject of movement or the instrument of action, it becomes rather the developer [révélateur] of time” (xi). It reveals (révèle) the spatio-temporal, ipso post facto, being-in-time-ness of a substance that, in my opinion, can no longer adhere to its own origins a minute or two ago, and of which Heidegger touches upon this notion of contemporaneity as Befindlichkeit as affect as mood: “mood assails. It comes neither from ‘without’ nor from ‘within,’ but arises from being-in-the-world itself as a mode of that being” (129). Of course, neither Heidegger, nor Deleuze for that matter, go so far as to suggest that time and affect are interchangeable concepts, as I do, but the one implies the other through a change, I argue, often through a change in the perspective of an individual, and quite literally when it comes to the Splat of Bowen Street, which needs to be seen from just the right point of view, ideally from the drone’s original perspective brought to a relativist scale; otherwise, when shown from behind (from where there is no data), the two-dimensionality of the image betrays itself. Jeff Malpas might then go on to suggest that, from “Deleuze through to Peter Sloterdijk spatial ideas and images are constantly in play, and yet what is at issue is the very idea of space and the spatial … . With some notable exceptions, very few thinkers, no matter what the discipline, have given serious attention to the phenomenon of space, any more than to the phenomena of time and of place, but have tended instead to deal with various forms or modes of space” (226). Malpas, I think, is right to question that we, as much as we are thinkers as we are humans first of all, tend to designate the phenomenon of space in much the same way that we are predisposed to utilise the reality of those spaces as places as concepts; not just through our perception of somewhere like a forest as a form of natural resource, but our own perspective of a standing dead or dying river red gum as an unsightly, and thus, undesirable mode of tree that, in any traditionalist park or garden setting, would sooner be cut down than left to hollow out as habitat for local wildlife. Edward Soja also touches upon this notion of a third space that “is not confined solely to geographers, architects, urbanists and others for whom spatial thinking is a primary professional occupation” (3), but this idea of a “constantly shifting and changing milieu of ideas, events, appearances, and meanings” (2). I also think a part of the problem with such appearances stem from the fact that spaces are nonhuman (to begin with), such that spaces become places over a period of time, and that we affect those spaces into places. We bring the phenomenology of a time and a place to a space, in other words, because our appreciation of the one only goes so far as our understanding of the other anthropomorphically, epistemologically-speaking, not unlike Guy Debord’s description of a psychogeography “whether consciously organized or not, on the emotions and behavior of individuals” (8). Malpas himself only goes so far as “to draw the concept of space as it appears within geography, in particular, into the sphere of what I term “philosophical topog

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.