What does the Attentional Theory of Cinematic Continuity say about why match-action, eyeline and shot/reverse-shot cuts feel continuous, and what should an AI cutter check (Smith 2012)?

Problem: continuity rules are taught as conventions; this is a cognitive account of why they work. Conditions: cuts between shots of a scene. Method (source's argument): viewers match perceptual expectations across a cut by attention, without needing a coherent representation of the depicted space; cuts are timed so the next shot supplies what attention was preparing to reach (a gaze shift, a movement, a face). The paper treats match-action, matched exits/entrances, shot/reverse-shot, the 180 degree rule and point-of-view cuts as consequences, and says the cut is not hidden in a blink or saccade because cuts rarely coincide with such periods. It also notes viewers default to the screen centre when a subject leaves the frame and are captured by suddenly revealed bright objects. Mapping to AI shot design (inference, labelled): when cutting generated clips, (a) end a shot with the direction of an eyeline or movement the next shot continues, (b) place the subject near screen centre at the first frame of the new shot, (c) avoid a bright or high-contrast distractor in the first frames. Failure modes: an eyeline or exit that points the wrong way creates a mismatch the viewer must correct with a saccade; unmotivated cuts. Limitations: a theory paper with a few eye-tracked examples; the author's own scene analyses use a feature-film clip; the 'no coherent spatial model' claim is a position, not a settled fact. Version/date: Projections 6(1), 2012, author manuscript; fetched 2026-10-23. Curator qualifications: All cited documentation excerpts located on the live pages by curator GET; no video sources. Limitations: A theory essay; its empirical support is cited from other studies. The AI-design recommendations are my inference and have not been tested on generated footage. Evidence status: statements located in the cited sources by a correlated desk review (Claude); not executed; not independently reviewed.

{
  "citations": [
    {
      "url": "https://doi.org/10.3167/proj.2012.060102",
      "note": "cited at https://ualresearchonline.arts.ac.uk/id/eprint/21187/2/6679.pdf; Abstract; Section on cut timing, note 4"
    },
    {
      "url": "https://ualresearchonline.arts.ac.uk/id/eprint/21187/2/6679.pdf",
      "note": "observed URL (dated/alias form of the cited catalogue source; canonical identity retained)"
    }
  ]
}