Infant Visual Perception - kellmanlab.psych.ucla.edu

1 109 CHAPTER 3 Infant Visual PerceptionPHILIP J. KELLMAN and MARTHA E. ARTERBERRYTH EORIES OF PERCEPTUAL DEVELOPMENT110 The Constructivist View110 The Ecological View112 The Contemporary Situation in Perceptual Theory113 BASIC Visual SENSITIVITIES IN INFANCY114 Visual Acuity114 Contrast Sensitivity116 Orientation Sensitivity119 Pattern Discrimination119 Color Vision120 Motion Perception124 SPACE PERCEPTION126 Kinematic Information127 Stereoscopic Depth Perception129 Pictorial Depth Perception131 OBJECT PERCEPTION134 Multiple Tasks in Object Perception134 Edge Detection and Edge Classification135 Detection and Classification of Contour Junctions136 Boundary Assignment137 Perception of Object Unity137 Perception of Three-Dimensional Form142 Perception of Size145 FACE PERCEPTION146 Preference for Facelike Stimuli146 Perceiving Information about People through Faces148 Mechanisms of Face Perception149 CONCLUSION150 Levels of Analysis150 Hardwiringversus

2 Construction inVisual Development151 Future Directions152 REFERENCES152 How Visual Perception develops has long been a centralquestion in understanding psychological developmentgenerally. During its emergence as a separate disciplinein the late 1800s, psychology was focused primarily onhow human knowledge originates ( , Titchener, 1910;Wundt, 1862), an emphasis inherited from concerns inphilosophy. Much of the focus was on the relation be-tween sensation and Perception , especially in prevailing view, inherited from generations of em-piricist philosophers ( , Berkeley, 1709/1963; Hobbes,1651/1974; Hume, 1758/1999; Locke, 1690/1971), wasthat at birth, a human being experiences only meaning-Preparation of this chapter was supported in part by researchgrants R01 EY13518-01 from the National Eye Institute andREC 0231826 from the National Science Foundation to thank Heidi Vanyo for helpful assistance.

3 Less sensations. Coherent, meaningful, Visual realityemerges only through a protracted learning process inwhich Visual sensations become associated with eachother and with touch and action (Berkeley, 1709/1963).Through most of the twentieth century, even as psy-chology increasingly emphasized findings of empiricalresearch, this primarily philosophical view cast a longshadow. Its inf luence was so great as to be essentially aconsensus view of development. William James (1890)echoed its assumptions in his memorable pronouncementthat the world of the newborn is a blooming, buzzing,confusion. Modern developmental psychology, shapedgreatly by Piaget, incorporated the same ideas. AlthoughPiaget combined contributions of both maturation andlearning in his theories, his view of the starting points ofperception was standard empiricist ( , Piaget, 1952,1954).

4 He did place greater emphasis on action,rather110 Infant Visual Perceptionthan mere sensory associations, as the means by whichmeaningful reality emerges from initially basic story about early Perception and knowl-edge persisted, in part, because researchers lackedmethods for investigating these topics scientifically. Thearguments of Berkeley and others were primarily logicalones. Claims about the origins of knowledge in the asso-ciation of sensations initially came from theory andthought experiments. Later, a few experiments withadults were used to make inferences about aspects ofperception that might be based on learning ( , Wal-lach, 1976) or not so based ( , Gottschaldt, 1926).Finding a more direct window into Perception andknowledge of a young Infant seemed unlikely. As Riesen(1947, p. 107) put it: The study of innate Visual organi-zation in man is not open to direct observation in earlyinfancy, since a young baby is too helpless to responddifferentially to Visual excitation.

5 Inthetime since Riesen s (1947) observation, the sci-entific landscape in this area has changed entirely. Al-though the development of Visual Perception is among themostlong-standing and fundamental concerns in thefield, it is also an area that is conspicuous in terms of re-centandrapid progress. Beginning in the late 1950s, thedoor to progress has been the development of methods forstudying sensation, Perception , and knowledge in humaninfants. The results of scientific efforts, continuing to thepresent, have changed our conceptions of how perceptionbegins and develops. These changes, in turn, have gener-atedimportant implications about the early foundationsofcognitive, linguistic, and social this chapter, we consider current knowledge ofearly Visual Perception and its development. Besides de-scribing the origins and development of these perceptualcapabilities, we use them to illustrate general themes:the several levels of explanation required to understandperception; the roles of hardwired abilities, maturation,and learning in Perception ; and some of the methodsthat allow assessment of early Perception .

6 These themesall have broader relevance for cognitive and EORIES OFPERCEPTUAL DEVELOPMENTAs a backdrop for considering research in early vision,we describe two general theories of perceptual develop-ment. These serve as valuable reference points in under-standing how recent research has changed our concep-tions of how Perception Constructivist ViewThe term constructivismhere refers to the view that per-ceptual reality must be constructed through extendedlearning. Choosing one term to label this idea is effi-cient, but also unfortunate, as this set of ideas has manynames. In philosophy, this kind of account is most oftencalled empiricism,emphasizing the role of input fromexperience in forming Perception . If, as is usually thecase, associations among sensations are held to domi-nate perceptual development, the position may also beaptly labeled the earliest days of psy-chology as an independent discipline, the merging ofcurrent and remembered sensations to achieve objects inthe world was called structuralism (Titchener, 1910).

7 Helmholtz (1885/1925) is often credited with applyingthe label constructivismto the idea that sensations arecombined with previously learned information using un-conscious inference to achieve perceptual reality. Thispedigree, along with Piaget s later emphasis on inputsfrom the learner s actions in constructing reality, makeconstructivism perhaps the best term to characterizemodern versions of this view. Unfortunately, the termhas been used elsewhere with different shades of mean-ing. In considering issues in learning, developmental andeducational psychologists often contrast constructivismwith associationism, where constructivism emphasizesthe active contributions of the learner. Although a com-mon thread extends through the uses of constructivism,its use here will be confined to the notion that percep-tion is constructed from sensations and actions throughlearning.

8 Our primary concern in addressing perceptionis to consider, not particular modes of learning, butwhether basic perceptual abilities are learned at all. Forthis reason, and others, the verdict on constructivism inthis domain may differ from the fates of constructivismsin other studies of constructivist account of how Perception devel-ops is familiar to many. The key assumption is that atbirth, sensory systems function to produce only theircharacteristic sensations. Stimulation of the Visual sys-tem yields sensations of brightness and color, along withsome quality (a local sign ) correlating with a positionon the retina. Stimulation of the auditory system pro-duces loudnesses and pitches, and so on. Of course, per-Theories of Perceptual Development111ceptual reality consists not of disembodied colors andloudnesses, but of objects arranged in space, relationsamong them, and events, characterized by motion andchange within that space.

9 On the constructivist view ofperceptual development, all these commonplace occu-pants of adult perceptual reality anytangible, materialobject existing in the external world and, indeed, the ex-ternal spatial framework itself are hard-won construc-tions achieved by allows construction ofexternal reality is associative processes. Experiences ofvisual sensations coupled with touch, according toBerkeley (1709/1963), allow creation of the idea thatseen objects have substance. Connecting the muscularsensations of reaching with Visual sensations allows thecreation of depth and space. Sensations obtained fromone view of an object at a given time are associated bycontiguity in space and time, and by similarity. Sensa-tions obtained a moment later from another view maybecome associated with the previous ones. An object be-comes a structure of associated sensations stored inmemory.

10 In John Stuart Mill s memorable formulation,for the mind, an object consists of all the sensations itmight give us under various circumstances: An object is the permanent possibilities of sensation (Mill, 1865).For Piaget (1952, 1954), the account is similar, exceptthat voluntary actions,not just tactile and muscular sen-sations, become associated, making objects consist ini-tially of sensorimotor did this basic story of perceptual developmentattain such preeminent status in philosophy and psychol-ogy? The question is puzzling because the account wasnot based on scientific study in any meaningful way. Justto anticipate a different possibility, we might considerthe life of a mountain goat. Unlike a human baby, amountain goat is able to locomote soon after birth. Re-markably, the newborn mountain goat appears to per-ceive solid surfaces on which to walk and precipices toavoid.