Stephen E. Palmer - US grants

The long term goal of the proposed research is to understand how the human visual system perceives and encodes properties of objects, such as their orientation, position, and motion in space, relative to frames of reference. More specifically, the research tests the implication of a theoretical distinction that is proposed between two kinds of visual implications of a theoretical distinction that is proposed between two kinds of visual reference frames: "world-surrogate" frames and "hierarchical" frames. World-surrogate frames define global reference values for properties such as perceived upright and stationarity, whereas hierarchical frames are used to define such properties only locally in encoding the relation of one perceived object to another. Several major predictions are suggested by this distinction and will be tested. (1) The objective orientation and motion of world-surrogate frames should be systematically misperceived as more upright and motionless that they are, whereas the orientation and motion of hierarchical frames should be veridically perceived. (2) World-surrogate frames should be effective to the extent that they are like an environment (i.e., large, surrounding, and temporally stable) whereas hierarchical frames should be effective to the extent that they are similar to the target object (in color, size, temporal onset, motion, depth, and so on) because "belongingness" is proposed to be crucial to hierarchical frames. (3) Reference frames should not be determined simply by stimulus structure, but by how the observer perceives that stimulus.. (4) Transformations in hierarchical frames should be sufficient to overcome strong tendencies to see an object's orientation and shape as constant. These and several further questions about reference frames will be investigated using speed and accuracy measures in novel applications of several well-known perceptual paradigms, such as the rod- and-frame task (Asch & Witkin), memory for shape at an orientation (Rock), configural interference on orientation and shape discriminations (Palmer), induced motion (Duncker), and configural motion (Johansson). The research is important because it formulates and tests hypotheses that, if true, would change the way we understand orientation, shape, and motion perception. This understanding is crucial, in turn, to being able to diagnose and treat people with perceptual handicaps and perceptual learning disabilities, both of which have substantial impact on their mental health, self image, and ability to function in society.

9355034 Palmer Twenty members of Berkeley's Institute of Cognitive Studies propose an integrated graduate research training program in the interdisciplinary field of cognitive science with a focus on spatial cognition. We request funds to support a five-year training program for five graduate students to consist of both coursework and research on the theme of spatial cognition, including spatial perception, language, and reasoning. Students will we required to take a set of five basic lecture courses and a series of research seminars to ensure adequate background knowledge in computer science, linguistics, psychology, and biology on topics integral to research on spatial cognition. Research training will occur as the trainees engage in research on topics within a number of well-established projects at Berkeley on the theme of spatial cognition, including: (l) cross-linguistic studies of lexical semantics, (2) the nature of spatical and temporal metaphors, (3)the acquisition of spatial language by children in different cultures, (4) the "psychophysics" of spatial terms, (5) the biological basis of spatial cognition, (6) the representation of spatial knowledge, (7) computational models for understanding, learning, and producing spatial descriptions, and (8) biological aspects of spatial cognition in animals and humans. Both faculty and trainees will take part in a series of research seminars focusing on the results of these research projects. ***

DESCRIPTION (Adapted from the Applicant's Abstract): The broad, long- term objective of the proposed research is to investigate and understand the principles of perceptual organization. Ever since the seminal work of Gestalt psychologists in the 1920s, the topic of perceptual organization -- essentially, the nature of part-whole relations among perceptual elements -- has been recognized as central in perceptual theory, but it has been curiously resistant to integration with modern theory. The investigator is reexamining some of the basic assumptions of the Gestalt work and finds them wanting. Specifically, the well-known "laws of grouping" are widely believed to operate early in vision and to define perceptual units, but the investigator currently believes that both of these propositions are false. Instead, he is finding that grouping is determined by relatively late processes that operate at or after perceptual constancy. He also argues that logic dictates a very different organizational principle -- which he calls uniform connectedness -- to create the perceptual units on which grouping processes operate. In the investigator's view, uniform connectedness provides initial access to the part-whole hierarchy of perceptual organization at what he calls the entry level. He proposes to continue his preliminary studies testing empirical predictions based on these and related theoretical ideas. He plans to use both standard subjective report methods and several novel reaction-time techniques that provide quantitative measures of performance on objectively defined tasks, including the repetition discrimination task and the primed matching task.

The present research project aims to understand how people respond aesthetically to the spatial composition and color composition of simple, visually displayed images. Despite the importance of aesthetic evaluations in our everyday lives -- not only when we visit museums, but also when we buy clothes, choose movies, take vacations, decorate our homes, and landscape our yards -- surprisingly little is known scientifically about what color combinations and spatial arrangements people prefer. Most previous treatments have come from philosophers, artists, and art theorists, who take a prescriptive approach by stating what one should prefer aesthetically. In this project the principal investigator takes a scientific, descriptive approach by analyzing what people actually do prefer aesthetically and the variables that determine these aesthetic preferences. Early findings on spatial composition show, for example, that people prefer symmetrical views of an object (such as a front-view of a face) to be positioned at or near the center of the image, but asymmetrical views (such as a side-view of a face) to be positioned laterally off center so that it faces into the frame. Early findings on color composition show, for instance, that people generally prefer pairs of colors that contain similar (rather than contrasting) colors, but that a given color looks more aesthetically pleasing against a highly contrasting (rather than a similar) color.

The proposed research project provides a promising new bridge between art and science. It defines an interdisciplinary research endeavor -- called aesthetic science -- that uses well-defined behavioral methods to understand an aesthetic dimension of human experience that is distinct from the well-studied cognitive (knowledge-based) aspects and emotional (feeling-based) aspects of experience. It also defines new research methods aimed at understanding different patterns of aesthetic preference between individuals across different aesthetic domains. Do people who prefer unusual, highly contrasting color combinations, for example, also prefer unexpected spatial compositions? Given the pioneering nature of the proposed research on an under-studied topic, all manner of new discoveries are possible, both within the field of aesthetic science and linking it with knowledge in the existing fields of emotion and cognitive science.

Abstract

People show strong preferences for certain colors and spatial compositions in visually displayed images, preferences that pervasively influence a multitude of decisions, such as buying clothes, taking vacations, decorating homes, and landscaping yards. Yet, surprisingly little is known scientifically about what color combinations and spatial arrangements people like or why they like them. Many previous treatments have come from philosophers, artists, and business consultants. Dr. Stephen Palmer and his research team at University of California -- Berkeley take a scientific, descriptive approach by measuring what people prefer and determining what factors determine these preferences. Further, they are also discovering why people have those preferences.

An early discovery of this line of research, for example, is that people's average preference for a given color is largely determined by how much they like the objects that characteristically have that color. For instance, people generally like blues because they generally like clear skies, clean water, and most other similarly colored objects. In addition, people show strong and systematic tendencies to prefer objects that are positioned at or near the center of the frame, but offset horizontally so that they face into (rather than out of) the frame, and offset vertically so that their position within the frame reflects their typical position relative to a viewer (e.g., a flying eagle toward the top of the frame and a swimming stingray toward the bottom).

The investigators propose to extend these projects in important ways. In the domain of color preferences, they will test their "ecological valence theory" of color preference by determining whether it can account for cross-cultural differences, using the same methods they have developed to test the theory previously. They plan to use similar methods to find out whether the theory can also account for individual differences in color preferences. In the domain of spatial composition, they plan to test a theory of compositional symmetries and asymmetries based on what they call "affordance spaces" around objects.

The proposed research project provides promising bridges among science, art, and business, using rigorous scientific methods to understand human preferences that are distinct from well-studied cognitive (knowledge-based) processes. It also defines new research methods aimed at understanding different patterns of preference between individuals across different domains. Do people who prefer unusual, highly contrasting color combinations, for example, also prefer nonstandard spatial compositions and less conventional music? The answers to such questions have important implications for business decisions and consumer marketing techniques. Given the pioneering nature of the proposed research on an under-studied topic, many new discoveries are being made that provide links between existing scientific knowledge in vision and cognitive science on the one hand and practical issues in the business world on the other.