The predictive-affective loop for directors: a target-state design framework for audiovisual directing.

In this Hypothesis and Theory article, we propose the PAL-D (Predictive-Affective Loop for Directors) framework as a mid-level model linking director-side design decisions to the viewer's predictive-affective response in film and television. Building on predictive processing, event segmentation research, and research on emotion and appraisal, we first lay out the viewer-side PAL (Predictive-Affective Loop) model, which characterizes viewing as a hierarchical Gaze-Prediction-Appraisal-Integration architecture in which prediction error, physiological arousal, appraisal, the prevailing affective condition, and an integrated episode-level affective condition interact over time. We then introduce PAL-D as a director-side grid that proposes four representative target-state patterns-calm, tension, crisis, and relief-four higher-order design axes (sensory density, framing and focus, narrative-contextual cueing, and affective-empathic cueing), and six perceptual-formal levers (camera position and distance, camera movement, framing and composition, light and color, editing rhythm and transitions, and sound and music). PAL-D treats the time course of prediction error as the primary organizing dimension and treats each target state as an intended trajectory of prediction error with characteristic arousal and affective profiles, against which empirically observed trajectories can later be compared. Rather than a fully quantified model, PAL-D is offered as a hypothesis-generating framework that makes explicit how specific combinations of target state, design axes, and formal levers are expected to modulate the viewer's predictive-affective loop. We outline applications to scene design, film and media education, experimental and observational research, and human-AI collaborative directing, and we indicate how its central quantities can be measured. PAL-D introduces no new psychological principles; its contribution is an integrative, predictive structure rather than a new mechanism. It is a mid-level model specialized for passive viewing, and it treats story-level construction, production conditions, and individual and cultural differences as wider-scope factors, isolating the segment-level predictive-affective trajectory through which these factors act. The framework thus aims to provide a common predictive-affective vocabulary through which filmmakers, media psychologists, and cognitive scientists can jointly analyze and test director-side design of visual narratives.