Conventions of Viewpoint Coherence in Film

1. All authors contributed equally to this work. The film sequences were produced with assistance of Judy Phu (Cinematography), Jeanne Tyson (Cinematography), Dylan Chapgier (Cinematography) , Nathan Meade (Assistant Camera/Gaffer), David Pregerson (Production Assistant), Jonathan Oldham (Actor), and David Atlas (Actor). We are grateful for the comments and suggestions of reviewers and colleagues, as well as audiences at the University of Oslo, Cornell, SCSMI, UCLA, and Haverford. This project was supported in part by the UCLA Office of the Vice Chancellor for Research Transdisciplinary Seed Grants. Over the last century, the philosophical study of meaning, including the areas of semantics and pragmatics, has overwhelmingly focused on language. In this paper, we examine the interplay of semantics and pragmatics in the alternative domain of film — understood to include all variety of video and animation, as they appear in cinema, on television, and online. Far from a marginal case, film is one of the primary means by which humans exchange narrative in contemporary society. A film consists of one or more shots organized into linear temporal sequence; our focus are those films which combine multiple shots to express their content. While, in principle, each shot in a given sequence could depict a completely isolated situation with no implied connection to the next, filmmakers have discovered methods of montage, or editing, by which shots are arranged so as to reliably convey unbroken stories, taking place in continuous space and time. How sequences of distinct images (and sounds) are able to express such coherent content is one of the central questions of classical film theory, and is the subject of this paper. The semantic view of film holds that the expression of unified content through the juxtaposition of shots is made possible in part by a system of semantic conventions. Like the rules of language, such conventions constitute specific interpretive principles which go beyond our everyday capacities for perception, thought, and action. In the case of film, they function to introduce implicit temporal, spatial, and narrative relationships between the events represented by adjacent shots. But, unlike some of the examples familiar from language, these conventions need not be fully arbitrary; instead, they may be naturally motivated in part by prior expectations and tendencies. For example, edited sequences following the common point-of-view convention move from a shot of a character’s face to a shot of the scene they perceive, echoing the natural transition from noticing a glance to looking at the