How do we come to know which elements of a visual display belong together as one grouping and which belong to different groupings? As the question was posed by Kohler1 when referring to Fig. 1, “… the reader has before him two definite groups of patches. Why not merely six patches? Or two other groups? Or three groups of two members each?” What Kohler1 and others from the Gestalt tradition2,3 have answered in response to this question is that perceptual organization “is an original sensory fact” because our perceptual systems are constrained to follow certain grouping principles that operate on the basis of the common movement, good continuation, proximity, and similarity properties of the elements. However, in response to the Gestaltists, Hebb4 adopted a constructivist stance and argued that even “simple perceptions” are “slowly and painfully learned” because eye movements that occur as one scans a visual pattern are needed to define the internal representation of the pattern. Such scanning is initially incomplete and requires both experience and the maturation of eye movements during real developmental time to construct a coherent mental record of a spatial arrangement of elements.
In the historical time period in which the nativist Gestalt and empiricist Hebbian positions were advanced, looking-time and eye-tracking methodologies were not yet available for studying perceptual organization in infants. However, such methodologies did become available during the 1960s and 70s and continue as mainstays in the toolkit of the modern infant investigator. Moreover, the findings that have emerged from the application of such methodologies have provided support for both the Gestalt and Hebbian views by demonstrating that a substantial degree of perceptual organization is present in the initial months of life, but at the same time, learning processes play a role in the acquisition of at least some aspects of perceptual organization. This article now turns to a review of the evidence.
Initial Eye-Tracking and Looking-Time Studies
Salapatek5 presented young infants with simple two-dimensional outline figures (i.e., square, triangle) and reported that scanning that was initially limited to single features became somewhat more extensive, although still incomplete, during the first months of life. Salapatek5 concluded that “features and their arrangement appear to be learned piecemeal and slowly” (p. 227). However, because of the difficulty of specifying the relationship between scanning and processing (i.e., processing may extend beyond the portion of the stimulus fixated by the infant), this conclusion was posed tentatively. As observed by Salapatek,5 “because of the difficulties in conducting habituation and conditioning studies with infants, we cannot be sure we have motivated the infant to tell us all that he perceives” (p. 232).
In one of the initial looking-time investigations of perceptual organization in infants, Spelke6 and Kellman and Spelke7 used a habituation procedure to ask whether 4-month-olds would perceive the continuity of a rod that was partially occluded by a block. On the basis of these studies, Spelke6 proposed an account of the development of object perception that may be thought of as a hybrid model incorporating Gestalt notions of innate or core organizing principles as well as a role for Hebbian learning based on experience with a structured environment. Specifically, Spelke6 argued that the infant’s perceptual system is constrained to follow two core organizational principles: common movement and connected surface. Adherence to these principles allows infants to parse from a visual scene a set of object blobs that maintain their coherence as they move. This experience, in turn, allows infants to register that such blobs tend to display similarity of surface and good continuation of contour. Thus, by the Spelke6 account, some of the original organizational principles identified by the Gestaltists (i.e., similarity, continuity) are actually “scaffolded” or “bootstrapped” onto the core principles by virtue of their correlation with those core principles.
Subsequent Looking-Time Studies of Infants’ Use of Classic Organizational Principles
Quinn et al.8 have undertaken a program of research to examine how young infants would utilize the classic Gestalt grouping principles to organize a variety of visual displays. The initial study asked whether 3-month-olds could utilize lightness similarity to organize visual pattern information.9 As depicted in Fig. 2, the infants were familiarized with stimuli consisting of columns or rows of elements that could be organized on the basis of their lightness vs. darkness, and the test stimuli were horizontal vs. vertical bars. The infants preferred the novel organization of bars, suggesting that they had grouped the individual elements of the arrays into rows or columns based on lightness similarity. Additional investigations conducted with similar methodologies showed that both good continuation and proximity are also functional at 3 to 4 mo of age.10,11
By contrast, when form similarity was investigated with the X-O contrast depicted in Fig. 3, 3- to 4-month-olds did not prefer the novel organization of bars even when provided with additional study time, although 6- to 7-month-old did show this preference.12 Moreover, the difficulty with form similarity at 3 to 4 mo also occurs for square-diamond and H-I contrasts.13 These results indicate a developmental change from 3 to 4 mo to 6 to 7 mo of age in sensitivity to form similarity as an organizational principle. Quinn and Bhatt13 also found that perceptual learning underlies this developmental change: when younger infants were “instructed” by providing three shape contrasts (X-O, square-diamond, and H-I) depicting a common organization during the same familiarization session, they did exhibit evidence of perceptual organization by displaying a preference for the novel organization of bars. Thus, learning based on variability in the experienced cues enabled younger infants to organize, like older infants, on the basis of form similarity.
Taken together, the findings on lightness similarity, good continuation, and proximity on the one hand, and form similarity on the other hand, indicate that different Gestalt principles become functional over different time courses of development and that not all principles are readily deployed in the manner proposed by Gestalt psychologists (see also Quinn and Bhatt14). The findings on form similarity are also consistent with the Spelke6 view that learning may play a role in acquiring at least some aspects of perceptual organization.
Infants’ Use of Newer Grouping Principles
Although the classic grouping principles have been known to psychologists since the 1920s, Rock and Palmer15 introduced the principle of uniform connectedness (UC) to refer to the visual system’s tendency to group together connected elements. To examine whether UC is functional in early infancy, as shown in Fig. 4, 3- to 4-month-olds were habituated to the uniformly connected patterns shown in Panels A or B, and then with a test pairing of connected vs. disconnected elements.16 A novelty preference for the disconnected-element test stimulus was observed, indicating that young infants are sensitive to UC. It is the contention of Palmer and Rock17 that UC allows for the delineation of entry-level units that can then be organized into holistic percepts according to other classic grouping principles (e.g., proximity, similarity). By this reasoning, the Hayden et al. data imply that 3- to 4-month-olds are able to use UC to identify the entry level units that are then subject to further perceptual organization.
A second newer grouping principle investigated was that of common region.18,19 This principle proposes that elements within a region are grouped together and separated from those in other regions.20 To investigate whether young infants could use common region to organize visual pattern information, 3-to 4-month-olds were familiarized to images with two pairs of shapes, with one pair (say A and B) always located together in a region and the other pair (say C and D) located together in another region, see Fig. 5.18 Infants were then tested with a familiar within-region grouping (e.g., AB) vs. a novel between-region grouping (e.g., BC). The infants discriminated between the grouping of elements from different regions and the grouping of elements that had shared a common region during habituation. These results indicate that enclosed shapes can be used by young infants to group elements within the shapes on the basis of common region. The data are also noteworthy given that the infants were habituated to vertical regions and tested with novel horizontal regions. When elements that were previously grouped based on a set of regions are subsequently encountered in a novel regional configuration, the infant perceptual system seems to expect this grouping to be intact. The findings are thus suggestive of a “unitization” process by which previously disparate elements become grouped and begin to function as coherent units in new contexts.
Top-Down Contributions to Perceptual Organization in Infancy
The review has thus far highlighted bottom-up configural determinants of perceptual organization in infants (e.g., common motion, similarity, connectedness). However, Needham and co-workers21–23 have emphasized that knowledge of particular objects and object kinds, reflecting top-down information, may also contribute to the development of perceptual organization. Such top-down information may be experientially acquired in the context of the organizational task itself or it may reflect a priori knowledge (i.e., innate knowledge or knowledge acquired outside of the context of the organizational task and prior to participation in it). A framework for thinking about the relative contributions of bottom-up perceptual process and top-down knowledge access to perceptual organization has been offered by Schyns et al.24 Schyns et al.24 argue that an individual’s history of concept formation (i.e., the concepts possessed by an individual at a specific point in time) will affect their subsequent perceptual organization processes.
Quinn and Schyns25 undertook experiments to determine whether the proposed interplay between adherence to Gestalt organizational principles and concept formation would be manifested in young infants. They examined whether features that are specified as coherent by Gestalt principles are “overlooked” by young infants if alternative means of perceptual organization are “suggested” by presenting the infants with a category of objects in which the features uniting the objects are “non-natural” in the Gestalt sense. As can be seen in Fig. 6, in an initial experiment, 3- to 4-month-olds were familiarized with a number of complex figures and subsequently administered a novelty preference test that paired a pacman shape with a circle. The infants were found to recognize the circle as familiar as evidenced by their novelty preference for the pacman. This finding suggests that the infants had parsed the circle from the complex figures in accord with good continuation (see also Quinn et al.26).
In follow-up experiments, Quinn and Schyns25,27 asked whether an invariant part abstracted during category learning would interfere with perceptual organization achieved by good continuation. The experiments consisted of two parts. In Part 1, the infants were familiarized with multiple exemplars, each marked by an invariant pacman shape and subsequently tested with the pacman and the circle (Fig. 7). The pacman was recognized as familiar, as evidenced by a preference for the circle. Part 2 was then administered and it followed the design of experiment 1 (Fig. 6). If the representation of the pacman carries over from Part 1 to Part 2, then infants should continue to prefer the circle (instead of preferring the pacman as had been the case in experiment 1). The latter result is what occurred and it suggests that the bias set by the Gestalt principle of good continuation may be soft-wired in the sense that it is subject to interference by processes of concept learning. More generally, an individual’s history of categorization will affect their subsequent perceptual organization of visual pattern information.
The research reviewed indicates that both classic and modern organizational mechanisms are available in early infancy, such that even infants as young as 3 mo of age are able to readily utilize a variety of bottom-up, stimulus-based cues to organize visual patterns (i.e., common motion, common region, connectedness, continuity, lightness similarity, proximity). However, other organizational processes appear to involve perceptual learning (i.e., form similarity). In addition, particular kinds of experiences (i.e., concept learning) may modulate the activity of grouping mechanisms. In infancy, then, as in adulthood, perceptual organization appears to be a function of both bottom-up and top-down processes. One likely fruitful direction for research will thus be to investigate the nature of perceptual learning and other top-down processes that affect visual organization and its development in infancy.
We thank Velma Dobson for comments on an earlier version of the manuscript.
Paul C. Quinn
Department of Psychology
University of Delaware
Newark, Delaware 19716
1. Kohler W. Gestalt Psychology. New York: Horace Liveright; 1929.
2. Koffka K. Principles of Gestalt Psychology. New York: Harcourt, Brace and Co.; 1935.
3. Wertheimer M. Principles of perceptual organization. In: Beardslee DC, Wertheimer M, eds. Readings in Perception. Princeton, NJ: Van Nostrand; 1958:115–35.
4. Hebb DO. The Organization of Behavior: A Neuropsychological Theory. New York: John Wiley and Sons; 1949.
5. Salapatek P. Pattern perception in early infancy. In: Cohen LB, Salapatek P, eds. Infant Perception: From Sensation to Cognition, vol. 1: Basic Visual Processes. New York: Academic Press; 1975:133–248.
6. Spelke ES. Perceptual knowledge of objects in infancy. In: Mehler J, Walker ECT, Garrett M, eds. Perspectives on Mental Representation. Hillsdale, NJ; L. Erlbaum Assoc; 1982:409–30.
7. Kellman PJ, Spelke ES. Perception of partly occluded objects in infancy. Cognit Psychol 1983;15:483–524.
8. Quinn PC, Bhatt RS, Hayden A. What goes with what? Development of perceptual grouping in infancy. In: Ross BH, ed. Psychology of Learning and Motivation, vol. 49. New York: Academic Press; 2008:106–46.
9. Quinn PC, Burke S, Rush A. Part-whole perception in early infancy: Evidence for perceptual grouping produced by lightness similarity. Infant Behav Dev 1993;16:19–42.
10. Quinn PC, Bhatt RS. Good continuation affects discrimination of visual pattern information in young infants. Percept Psychophys 2005;67:1171–6.
11. Quinn PC, Bhatt RS, Hayden A. Young infants readily use proximity to organize visual pattern information. Acta Psychol (Amst) 2008;127:289–98.
12. Quinn PC, Bhatt RS, Brush D, Grimes A, Sharpnack H. Development of form similarity as a Gestalt grouping principle in infancy. Psychol Sci 2002;13:320–8.
13. Quinn PC, Bhatt RS. Learning perceptual organization in infancy. Psychol Sci 2005;16:511–15.
14. Quinn PC, Bhatt RS. Are some gestalt principles deployed more readily than others during early development? The case of lightness versus form similarity. J Exp Psychol Hum Percept Perform 2006;32:1221–30.
15. Rock I, Palmer S. The legacy of Gestalt psychology. Sci Am 1990;263:84–90.
16. Hayden A, Bhatt RS, Quinn PC. Infants’ sensitivity to uniform connectedness as a cue for perceptual organization. Psychonom Bull Rev 2006;13:257–61.
17. Palmer S, Rock I. Rethinking perceptual organization: the role of uniform connectedness. Psychonom Bull Rev 1994;1:29–55.
18. Bhatt RS, Hayden A, Quinn PC. Perceptual organization based on common region in infancy. Infancy 2007;12:147–68.
19. Hayden A, Bhatt RS, Quinn PC. Perceptual organization based on illusory regions in infancy. Psychon Bull Rev 2008;15:443–7.
20. Palmer SE. Common region: a new principle of perceptual grouping. Cognit Psychol 1992;24:436–47.
21. Needham A. Object recognition and object segregation in 4.5-month-old infants. J Exp Child Psychol 2001;78:3–22.
22. Needham A, Baillargeon R. Effects of prior experience on 4.5-month-old infants’ object segregation. Infant Behav Dev 1998;21:1–24.
23. Needham A, Dueker G, Lockhead G. Infants’ formation and use of categories to segregate objects. Cognition 2005;94:215–40.
24. Schyns PG, Goldstone RL, Thibaut JP. The development of features in object concepts. Behav Brain Sci 1998;21:1–17.
25. Quinn PC, Schyns PG. What goes up may come down: perceptual process and knowledge access in the organization of complex visual patterns by young infants. Cogn Sci 2003;27:923–35.
26. Quinn PC, Brown CR, Streppa ML. Perceptual organization of complex visual configurations by young infants. Infant Behav Dev 1997;20:35–46.
27. Quinn PC, Schyns PG, Goldstone RL. The interplay between perceptual organization and categorization in the representation of complex visual patterns by young infants. J Exp Child Psychol 2006;95:117–27.