Deictic Gaze

Historical reception studies of the sort just described are concerned with economical, productional, (non)narrative, and stylistical contexts within which conventions and discursive practices are used and understood by its audience. To a psychologist, this is vital information. To describe the mental processes of the spectator, however, we must now begin to sketch the mental dispositions involved and how they influence the understanding of POV editing. My general approach is that understanding POV editing is fundamentally an inferential activity in the cognitive unconscious of the spectator. Whether in cinema or in real life, the spectator infers that a person is looking at a certain object. This inference is never objectively certain, but only more or less probable and always relative to understanding (like all cognition and inferences). Even when the gazer and the seen object are in the same shot, the spectator still must attribute a gaze (and the object of that gaze) to the character. C looks at O or This is what C looks at is thus never a textual structure per se, but a mental coherence created on the basis of textual cues and, as I argue, deictic-gaze ability. Some scholars have acknowledged this inferential feature of POV. Thompson (1985:203) uses terms such as surmise: a character looking offscreen in one direction might lead the viewer to surmise that the next shot shows the space that Understanding Point-of-View Editing 67 character sees. Jean Mitry talks about the processes of recognition:
In any case, the subjective image is never more than a complement to another image. It has meaning only insofar as it relates to a character already objectively described and placed. I can see what Pierre sees only if I have already seen Pierre, and I can share his point of view only if I can relate it to him, recognizing it as his. (Mitry, 1997:209)

Although both of these passages indicate the importance of understanding processes in the spectator, none of them goes into detail. My emphasis on understanding rather than on textual structure contrasts with semiotic structuralist approaches, which deal more or less exclusively with textual processes and the reader that the text implies or suggests (Branigan, 1984). A psychological approach does not deny that textual structures exist, but their shape and effects cannot be deduced solely on the basis of textual analysis. A description of POV editing must be complemented by an investigation into the interaction among spectators biological, psychological, cultural, and historical dispositions and the text. Whereas cultural reception theorists grasp such dispositions through analyzing cultural phenomena and discourse, I make use of psychological studies. So how, then, is the relation between glance shot and object shot recognized, surmised, inferred, or understood by the spectator? Carroll (1993) presents a framework to explain these processes. The spectator understands POV editing, Carroll maintains, because it is a representation of an event that is basic to humans, namely, the habit of following and determining the object of another persons gaze, so-called deictic gaze. [P]oint-ofview editing [. . .] works because it relies on depicting biologically innate information-gathering procedures (Carroll, 1996a:129). Although Carroll is reluctant to go into details of the mental processes of understanding an endeavor he may gladly pass on to psychologists his basic idea is worth exploring, as it presents a way not only to describe mental processes of understanding, but also, as we will see, to explain stylistic features and changes of the POV convention. To this end, we must take a closer look at what deictic gaze is and how it is structured.

The Structure of Deictic-Gaze Behavior

Gazing in humans involves a wide set of behaviors with a similarly wide set of functions, for example, social, emotional, and information gathering (Kleinke, 1986). A simple distinction, though, can be made between mutual gaze, in which two persons look at each other, and deictic gaze,

in which person A monitors the direction of person Bs gaze and tries to establish the target of this gaze. Deixis is a linguistic term, denoting 68 Understanding Cinema a reference by means of an expression whose interpretation is relative to the (usually) extra-linguistic context of the utterance, such as who is speaking, the time or place of speaking, the gestures of the speaker, or the current location in the discourse (Linguistic Glossary, 1997). English examples would be I, You, Now, There, That, and The following. Deictic gaze thus works as pointing and other indexical practices, which contextualize and anchor linguistic utterance in space. Deictic gaze creates a perceptual space common to both participants and is thus often referred to as joint visual attention (Butterworth, 1991). It is primarily this latter form of gazing behavior that I discuss. development. Because deictic gazing is a basic form of human behavior, it is easily overlooked and taken for granted. To understand its fundamental importance in human life and to emphasize that it is an acquired ability, psychologists have turned to toddlers. In the typical experimental setting, infant and parent are placed in a room with different target objects, e.g. toys. The parent is instructed to interact naturally with the child and then, at a signal, to turn, silently and without pointing, to inspect one of the objects. The reaction of the infant is recorded on video and judged by two independent observers, who estimate the direction and the accuracy of the infants response relative to the objects of the parents line of gaze. Results suggest that deictic-gaze competence develops in three stages, with each stage introducing new abilities (Butterworth, 1991). At the age of six months, babies apprehend and are quite sensitive to the change in the parents gazing behavior. They make clear efforts to investigate the direction of the parents attention. They also look at the correct side of the room for the target. When there are two identical targets on the same side of the room, however, the six-month-old babies cannot differentiate between them, although the objects may be separated as much as 60 degrees. When the correct target is the first one on the scanning path from parent to target, babies are mostly accurate, but when the correct target is second on this path, they perform only at chance level; that is, these children cannot determine the right target on the basis of the parents action alone, but the perceptual salience of the object and its setting seem to play a part (e.g., differentiating color, movement, and shape). At six months of age, the babies also fail to follow gaze if the parent looks at an object outside the field of view, for instance, behind the baby. These findings cannot be attributed to an inability to turn around, as children at this age often react that way to noises and other events. Instead, Butterworth (1991) suggests, the most likely explanation is that the child lacks an awareness that it is surrounded by a continuous space. The adults Understanding Point-of-View Editing 69 expectation of a space existing outside the immediate field of view appears to be absent in the six-month-old child. This corresponds well to Piagets theory about object and space permanence (recapitulated and scrutinized by Harris, 1983:715ff), a stance that a baby supposedly develops after the first year of life. Irrespective of the fact that six-month-old babies inability to look for objects behind them can be attributed to a lack of space permanence, it is fair to assume that, for deictic gaze to function properly in adults, this spatial assumption has to be in place. By the age of twelve months, the infant is beginning to localize targets correctly, whether first or second along the scan path. This suggests that some form of geometric ability is now definitely in place, independent of the perceptual salience of the target object. Butterworth observed in the experiments how the infant fixated on the mother while she was turning;

only a second or so after her turning ended did the child turn its head to the target objects, indicating that some form of angular estimation process occurred. It is interesting to note that this geometric ability surfaces at the same time as index-finger pointing starts to be understood by the child (toward the end of the first year, a couple of months before production of the gesture). Both abilities seem to include the understanding of a projection or extension of a straight line in space7 (perhaps connected to the act of throwing things and monitoring their trajectory). Still, however, twelvemonthold children failed to search for targets located behind them. Even though the visual field was emptied of targets, the children scanned only approximately 40 degrees of visual angle; when they encountered no object, they gave up the search (Butterworth, 1991:227). By the time a child is eighteen months of age, geometrical ability is refined, although not complete. Adults can discriminate changes of visual angle as small as 1 minute of arc of visual angle (Bruce,Green&Georgeson, 1996:357). That ability is probably not in place at eighteen months of age. Moreover, an infant is reluctant to search behind itself unless the space in front contains no objects at all. Although the assumption of object permanence is gaining force in the childs spatial understanding, it is not as developed as that of adults. Among children and adults, deictic-gaze behavior seems triggered by some sort of change in the gazers behavior: The gazer makes a clear shift of direction of gaze, becomes quiet or still, or changes from a wandering gaze to a still one. When Butterworths parents not only gazed but also pointed with arm and index finger, deictic-gaze behavior was triggered to a significantly larger extent in all age groups than with gaze only (Grover, 1988; retold in Butterworth, 1991). This indicates that some gazer-related event prototypically must be acknowledged in order to trigger deictic gaze behavior. 70 Understanding Cinema functions. Deictic gaze fulfills important social functions in humans and animals. First, joint visual attention creates among the participants a common semanticperceptual space, which enables and supports communication and language learning. Caretakers, for instance, are exceptionally sensitive to the direction of a childs gaze in order to establish what object, event, or situation catches the attention of the toddler at the moment. It has been shown that caretakers adapt and constraint their language use in order to level with the toddler and the object of attention, for instance, in labeling objects, activities, or situations (Bruner, 1983:67ff). Fromthe childs perspective, this shared perceptual space makes language learning much easier and faster, as the reference of the adults discourse is right there in front of its eyes. In adults, shared visual attention often has crucial functions in everyday discourse. Indexical gazing and pointing provide the context in which verbal utterances or other behaviors are meaningful and understandable (the fluctuating reference of this and that