Archives, recommendations, and viewing feedback
How do old images change meaning after preservation? How do platforms turn recent viewing into a ranking condition for the next item? Start with Deleuze's crystal, virtual, and sensory-motor schema, then follow the atlas's two research proposals into film analysis, counterexamples, and filmmaking exercises.
How does the crystal's temporal relation enter the archive?
Deleuze's crystal-image holds actual and virtual distinct while they exchange. The atlas's archival proposal follows this temporal relation into the custody, opening, and re-editing of images after filming, asking how later knowledge changes their historical meaning. The crystal describes a temporal relation within the image; the archival proposal examines uses of material at different historical moments.
Conceptual source: Deleuze's account of the exchange between actual and virtual. Archival extension: the atlas's future-perfect archive proposal. The questions of custody and use belong to this later research proposal.
How does recommendation probability differ from the virtual?
A recommender model computes probabilities for candidate items; Deleuze's virtual is not a list of candidates that have not yet appeared. Platform research asks how viewing becomes a ranking signal and how that signal shapes what appears next. Probability describes the system's calculation over candidate items; the virtual concerns a different temporal relation.
Deleuze's sensory-motor schema examines how perception extends into action. Governance of nextness moves the question to viewing feedback: how pausing, skipping, and replaying affect subsequent images. Clicks and recommendations are contemporary objects of the atlas's analysis.
How can evidence change after material is preserved?
The future-perfect archive moves a temporal relation within memory-images into an archival chain: filming, naming, classifying, storing, opening, deleting, and re-editing. Later encounters with events, testimony, or other images can change what the material establishes. Later knowledge alters today's reading; what the filmmaker knew at the time remains a question for the material available then.
Renaming old footage, restricting access, or placing it beside new testimony can change its context and evidentiary relations. The footage's age and the force it later acquires as evidence are separate questions.
Meeting with Pol Pot: what do archive and re-enactment change?
The comparison in Meeting with Pol Pot asks how archive and re-enactment arrange provenance, historical absences, and the relations on which spectators base their judgements. If they create only a historical mood without changing provenance, absence, or evidence, the film cannot support the future-perfect archive.
Watch whose material the image and sound identify, which historical positions remain absent, and whether the order of archive and re-enactment changes the understanding of events. These relations determine the film's connection to the archival proposal.
How does fixed playback differ from feedback-driven ranking?
Governance of nextness examines a feedback relation in which pausing, skipping, or replaying enters the ranking of subsequent images. A fixed playlist differs from that feedback sequence. The proposition applies only where viewing feedback actually participates in ranking.
If checkable records show that a studied platform sequence is fixed and uses no viewing feedback, governance of nextness does not apply to that case.
How do sound, duration, and editing order change understanding?
Six proposed film exercises work with shot count, sound source, duration, action, and editing order. Changing one condition and comparing how actions connect, sound relates to the image, or past meets present can make a conceptual question visible.
For example, keep the images while changing the sound source, or keep the material while changing its editing order. Does the same temporal relation survive both viewings? The exercise page provides the filming constraints and comparison questions.
The archive's past and the platform's next moment
The future-perfect archive asks how naming, classification, storage, opening, and re-editing after filming allow later knowledge to change our understanding of the past.
Governance of nextness and the right to stop asks how recommenders arrange subsequent images through viewing feedback, and how viewers can interrupt that sequence. Both propositions develop their concepts, cases, and counterexamples.