This one-hour programme presents ten experimental shorts that recontextualise found or archival footage via algorithmic intervention. These works of “retracted cinema” depart from the well-established practices of “expanded cinema” in which film materials are recontextualised through the supplementation of and collision with additional materials and extrinsic media, such as live performance. By contrast, the works presented in the Retracted Cinema programme use the appropriated film material itself as the site and means of self-recontextualisation. The results are achieved by applying a set of rules or constraints – an algorithm – that governs a repeatable (iterative) array of transformations. In this way, the artistic conception forges a bridge linking the algorithmic impulses within the traditions of avant-garde cinema, literature and visual art. The work calls to mind the détournement of found footage (recycled cinema) and an “inverted” expanded cinema, OuLiPo (Ouvroir de littérature potentielle), and conceptual art (from Duchamp to Sol Lewitt, Hanne Darboven and beyond).
In the era of artificial intelligence and machine learning, data analytics and surveillance capitalism, systems run according to instruction-sets that largely escape the public’s visual detection while serving and expanding the interests of international capital. Yet alongside the instrumental logic of techno-economic enterprises, another logic is at work in the algorithmic that overlaps with the avant-garde.
The driving aesthetic of tonight’s programme resists instrumentality by foregrounding and conceptualizing the use of the algorithm. Evading conventional artistic conceits and maneuvers, the work clings neither to the uniqueness of a precious unrepeatable result nor to the authentic unfolding of a creative process. These artistic productions, like OuLiPo and conceptual art, underscore a radical potentiality in their concept. The algorithmic constraints that here render an audiovisual result do so by generating a work of potential art. The projects parse and renegotiate the expedient nexus between conception and execution that underlies instrumentality while bestowing upon the result (with its component parts) the status of an archive awaiting future creative interventions. And rather than rejecting the mechanistic character of technological operations, the inward self-referential complexity of the works embraces the “machinic” (digital coding, algorithmic thinking) as the crux and material of artistic conceptualization, and not simply the instrumental means of implementing ideas.
The filmmakers in the Retracted Cinema programme owe a debt to the structural films of the 1960s and 1970s. But this programme hopes to present more than a cross-section of the living legacy of structural filmmaking. It aims to reveal new interests that reflect back on a previous generation of artists in order to forge new ground beyond the utopian aims of their predecessors’ materialism.
Gonzalo Egurza’s 17-17 (2017) launches an algorithm that extracts the individual frames of the entire Odessa Steps sequence (Battleship Potemkin) like a deck of cards and then reshuffles them. Blanca Rego’s Psycho 60/98 (2016) investigates the intersection of replication and montage, producing an interruptive flow of frame-by-frame cross-cutting between two versions of Psycho, the original Hitchcock and Gus Van Sant’s remake. Albert Alcoz’s Home Movie Holes (2009) utilizes a systematic strategy to decompose fragments recuperated from amateur films of the family by combining and recombining them with material film artefacts taken from 8mm reels. The piece presents a nostalgic reexamination of the digitization of film material that reveals the fragility of memory. Gregg Biermann’s Happy Again (2006) reorganizes the signature scene from the Hollywood musical Singin’ in the Rain through a method of programmatic layering and displacement. The result uncovers a new cinema, music and dance buried within the familiar iconic sequence. Vitor Magalhães’ Naturalezas muertas (en seis movimientos) (2019-20) utilizes a programmatic approach based on the selection and juxtaposition of particular moments, taken from roughly seventy films, in which the actors are hidden from the camera. The film embodies a kind of photo-roman or photo-essay organized into a set of six distinct “rooms of thought” populated by the mood of objects and décors, interspersed with written formulations.
Some of the computational films directly embrace the elusive character of the automaton with an interest in the performative dimension of the machinic. Taller Estampa’s ¿Qué es lo que ves, YOLO9000? (2019) advances a heterodox audiovisual investigation into one of the many contemporary artificial vision tools being developed and applied for automatic image recognition and annotation: YOLO9000. Estampa’s work applies this trained neural network, containing a dataset of 9,418 words and millions of images, to appropriated selections from the world of 20th-century avant-garde film and archival footage with surprising results. Barbara Lattanzi’s Optical De-dramatization Engine (2015) presents an artist-designed software that remediates footage from Thomas Ince’s 1912 film, The Invaders, moving between the levels of image, frame, and pixel. With each new launch of Lattanzi’s software, which also manifests as a 40-hour installation loop, a different remediation of Ince’s film unfurls.
Peter Freund’s Floating Point (2020) activates an algorithm that re-presents and re-frames an iconic scene from Stanley Kubrick’s 2001: A Space Odyssey. The algorithm embeds the original sequence within a larger, hinted-at grid system that tabulates the sequence’s twenty-five shots. While presented in the flat, two-dimensional cinema screen, the structure of the grid forms an “impossible geometry” whose corners and edges have been folded back to meet each other in the manner of a conceptual origami. Eloi Puig’s ongoing Torvix series (2011-Present) performs a software remediation of appropriated public YouTube videos by deriving grammatological features in the transcribed vocal track of the original footage as a prompt for transposing the footage into a shortened montage that responds only to the position of each letter of the text that is heard. Finally, Kuku Sabzi’s Lost Footage (2020) mobilizes the predictive logic of text generation machine learning to regenerate scan-line by scan-line the 1967 Patterson-Gimlin film purportedly documenting a Bigfoot traversing a forest in Northern California.