-
17. Movies in the Mind's Eye
- University of Wisconsin Press
- Chapter
- Additional Information
17 Movies in the Mind's Eye Julian Hochber;g and Vir;ginia Brooks Most writers on film, and most filmmakers, need no science. But any serious discussion of whether the medium was used effectively or artistically in any instance requires some understanding of how we perceive and remember moving pictures, and that must derive from research: introspection will not serve. Scattered aspects ofcognitive science have begun to appear, therefore, in recent writings on film.l On the other side ofthe aisle, students ofperception and visual memory cannot afford to ignore moving pictures, but until recently they have mostly confined their attention to low-level motion phenomena , as have introductory film texts. The latter, if they write at all about perception, still proclaim that we perceive motion from successive still frames because of(heavens protect us!) "persistence ofvision." 2 In any case, stroboscopic motion is only a small part ofthe perception ofvisual events, which is what film is about. In this essay we reexamine the cognitive systems that contribute to the visually informative and artistically important characteristics offilm and tape, trying to keep both the science and the art in view. Depicting Events in Moving Pictures There are three steps to the depiction ofevents in moving pictures: low-level vision, relational parsing, and action schemas. We will look at each separately. Movement as Primitive Sensory Response A continuous motion in the world is, of course, captured by successive displaced images on film (or their video equivalent). For most events, these displacements are small, and within the range of the low-level sensory receptors of the visual system; these respond identically to the small displacements on the screen and to the differences provided from one moment to the next by 368 HOCHBERG AND BROOKS: Movies in the Mind)s Eye 369 smooth physical motion in the world. Recent studies of this system offer an increasingly important window on the underlying neurophysiology, and reveal some surprising phenomena (for example, reversals of direction).3 This low-level response is preemptive and "unintelligent." Notably, it occurs between nearby successive contours regardless of what objects they belong to (see, for example, Fig. 17.3A), thereby causing many "bad cuts" (which result from this fundamental displacement-detection mechanism, rather than some violation of cinematic "grammaticality"). Many techniques aimed at achieving "seamless" editing work by avoiding such unwanted apparent motion between noncorresponding objects. In the time ofEisenstein, however, and especially in the New Wave, such visual jolts became desirable, although (or because) they slow the viewer's comprehension and make the medium itselfmore intrusive. Although vision psychologists and neurophysiologists sometimes write as though these low-level mechanisms account "directly" for perceiving motion , which would, if it were true, make it easy to explain and predict what people will see, it is simply not true: we usually perceive movements very different from the displacements in the eye or on the screen. Indeed, were it not for these differences, films as we know them would not be possible, as we see next. Framework-Relative Paths of Motion We perceive (approximately) the framework-relative paths ofmotion, and not the displacements on the screen which determine low-level motion. An object may be perfectly stationary on the screen and yet it will appear irresistibly to move ifgiven a moving framework (Fig. 17.1Ai) or background (Fig. 17.1Aii), and the actual motions of the frameworks or backgrounds themselves are often not noticeable. This is part ofa rich body ofphenomena known as induced motion.4 Something akin to Fig. 17.1Ai happens outside of the laboratory when the eye tracks a moving object in a pursuit movement or when a camera acts similarly in a pan or track shot (Fig. 17.1Aiii).5 Thanks to this phenomenon, a continuous motion can be presented over a space s' that may be many times larger than the screen, in the same way that the movements of the viewer's head and eyes provide a wider prospect than the limits ofgaze within any glance. The screen would be stage-bound, were it not for this resource. Similarly, the parts of an object or group of objects moving in one direction on the screen may instead be irresistibly seen as moving in another (Fig. 17.1Bi, 17.1Bii). Neither of these demonstrations is only about dots in the laboratory. The phenomenon of Fig. 17.1A reappears in Fig. 17.1c: The rightward movement of a dancer across the screen (M2) is lost if filmed in limbo...