Good Continuation

When considering the relevance of the Quinn et al. (1993, 2002) demonstrations of infant Gestalt-like grouping via lightness and form similarity for understanding the early development of object recognition abilities more generally, two limitations become apparent. First, the stimuli used in the Quinn et al. studies are more like surface textures than they are like objects (Spelke, Breinlinger, Jacobson, & Phillips, 1993; Spelke, Gutheil, & Van de Walle, 1995). Second, in many natural scenes, numerous objects appear simultaneously, and may be arranged so that their contours are partially overlapping, rather than completely visible. Such scenes must be parsed into a set of primitive contours, and these contours must be organized into coherent representations of whole shapes.

To examine how well infants parse and organize more complex visual configurations that only begin to approximate the object processing demands of more natural scenes, Quinn et al. (1997) investigated how young infants would represent the pattern shown in panel (a) of Figure 5.4. Adults tend to parse and organize this pattern into the square and circle shapes shown in panel (b) of Figure 5.4, rather



Figure 5.3 Familiarization and test stimuli used to test adherence to the form similarity Gestalt organizational principle in Quinn, Bhatt, Brush, Grimes, and Sharpnack (2002). The rationale is that if infants can organize the familiar stimulus in the top panel into columns, then the vertical-column test stimulus should be perceived as familiar, and the horizontal-row test stimulus should be preferred. Similarly, if infants can organize the familiar stimulus in the bottom panel into rows, then the horizontal-row test stimulus should be perceived as familiar, and the vertical-column test stimulus should be preferred.


Figure 5.4 Intersecting circle-square stimulus in (a) and two possible organizations in (b) and (c). Adherence to the Gestalt principle of good continuation favors organization (b) over organization (c). (From Quinn, Brown, & Streppa, 1997.)

than the pacman shapes shown in panel (c). Presumably adults are following Gestalt organizational principles such as good continuation to represent the pattern information in this manner.

To determine whether 3- and 4-month-olds would parse and organize the pattern shown in Figure 5.4a into the square and circle shapes perceived by adults, a group of infants was familiarized in an experiment with the overlapping square-circle configuration and then preference tested with the circle versus P1 and the square versus P2. There are at least three ways in which infants may represent the familiar pattern information, each of which would produce a distinct pattern of looking among the test stimuli. First, if infants parse the pattern and organize its contours into the circle and square shapes in accord with the Gestalt principle of good continuation, then the circle and square shapes should be recognized as familiar in preference tests with P1 and P2, and P1 and P2 should be preferred. Second, if infants do not parse the unitary configuration, and the outcome is a representation of an unparsed whole, or if infants parse the configuration into more than one whole, but not the wholes predicted by adherence to good continuation, then one would not expect a consistent preference for P1 or P2 in either test. Third, infants may have spontaneous preferences for certain stimulus features (e.g., curvature, horizontal/vertical line elements) that could eventuate in parsing of the configuration and selective organization of the features into either the circle or square. In this case, one would expect a preference for the pacman shape (P1 or P2) in one or the other preference test, but not both. The results were that infants preferred the pacman shapes in both preference tests, a pattern of looking that is consistent with the idea that infants had organized the familiar configuration into the circle and square shapes in a manner predicted by adherence to the Gestalt principle of good continuation.

A potentially informative effect that was observed in Quinn et al. (1997) is that there was a spontaneous preference observed in a control group of infants who were assessed for possible a priori preferences among the test patterns. Specifically, without prior familiarization experience, the P2 shape was significantly preferred to the square. Although the novelty preference for P2 over the square in the experimental group occurred above and beyond that of the a priori preference for P2 in the control group, it is interesting to consider the possible significance of the spontaneous preference for P2.

The critical physical difference between P2 and the square is the presence of a curved contour in P2 and the absence of it in the square. In addition, there is evidence that infants have a spontaneous preference for curved over rectilinear contours (Fantz, Fagan, & Miranda, 1975). It is conceivable that the a priori preference for P2 is more than just a potential confound, a noise factor to be acknowledged, but then ruled out with appropriate statistical tests. The curvature-based preference for the P2 stimulus may help answer the question: How does the infant initially begin to break down a complex stimulus into a set of component contours? More specifically, where on the stimulus does the infant begin the complementary processes of parsing and organization? It may be that spontaneous preferences for some stimulus features over others could play an important "START HERE" role in initiating the parsing process for an unanalyzed configuration that contains a number of diverse features. A curvature preference, combined with a Gestalt principle like good continuation, may have allowed infants to first develop a representation for the circle shape. Once the circle was organized, infant attention may then have been free to explore other portions of the stimulus, thereby allowing for organization of the remaining contours into the square shape.


The evidence reviewed in the preceding sections indicates that young infants can group together the elements of a single visual pattern, presented in isolation from other patterns, so as to form a holistic representation of that pattern. Young infants may also be capable of parsing and organizing the more complex pattern information in a configuration of intersecting contours into two complete shapes. In achieving this degree of perceptual coherence infants may benefit from adherence to certain Gestalt principles such as lightness similarity and good continuation at 3 to 4 months, with form similarity beginning to exert a contribution to organization at 6 to 7 months. In addition, infants may rely on spontaneous preferences for particular features of contour information such as curvature to begin parsing and organizing global conglomerations of visual pattern information that contain multiple features.

The data contrasting the developmental onset of infants' use of lightness versus form similarity challenge the traditional Gestalt claim that all organizational principles are automatically and equivalently applied (Kohler, 1929). They are also consistent with recent studies from the adult literature suggesting that not all Gestalt principles are equally powerful or operational at the same time in the overall course of processing (Behrmann & Kimchi, 2003; Peterson, 2001). The findings suggest further that young infants are sensitive to more than just common movement and connected surface principles as organizers of visual pattern information (Kellman, 1996; Spelke, 1982). In particular, the findings of Quinn et al. (1993, 1997) indicate that lightness similarity and good continuation information are functional as organizational principles as early as 3 months of age (and possibly earlier in the case of lightness similarity; Farroni et al., 2000). These data are thus convergent with the work of Johnson (1997) and Needham (2001) in suggesting that a range of cues may be operational as sources for organization, although the difference in the developmental emergence of lightness similarity versus form similarity is consistent with the idea that some cues for organization may carry more weight than others at different points during development. As such, investigators may need to think in terms of how differentially weighted cues come together, in either a threshold or integration framework, to give rise to one or another percept.


Earlier in the chapter, an important distinction was made between the contributions of perceptual process and knowledge access toward explaining the performance of infants participating in looking time studies. A framework for thinking about the roles of perceptual process and knowledge access in perceptual organization has been offered by Schyns, Goldstone, and Thibaut (1998). Schyns et al. argue for a flexible system of perceptual unit formation, one in which the features that come to define objects are extracted during the task of concept learning. The idea is 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 Schyns (2003) undertook a set of experiments to better understand the interplay between adherence to Gestalt organizational principles and flexible feature creation. The experiments were designed to answer the following question: Will features that are specified as functional by Gestalt principles be "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 "nonnatural" in the Gestalt sense? In the first experiment, 3- and 4-month-olds were

Familiarization Trials

Test Trials

Figure 5.5 Examples of the familiarization stimuli and test stimuli used in Quinn and Schyns (2003). If the infants can parse the circle from the familiar patterns in accord with good continuation, then they should prefer the pacman shape over the circle shape during the test trials.

familiarized with a number of complex figures, examples of which are shown in the top portion of Figure 5.5. Subsequently, during a novelty preference test, the infants were presented with the pacman shape paired with the circle shown in the bottom portion of Figure 5.5. The infants were found to recognize the circle as familiar as evidenced by their preference for the pacman shape. This finding is consistent with the idea that the infants had parsed the circle from the complex figures in accord with good continuation.

In follow-up experiments, Quinn and Schyns (2003) asked whether an invariant part abstracted during category learning would interfere with perceptual organization achieved by adherence to 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 were subsequently administered a novelty preference test that paired the pacman shape with the circle shape. Examples of the stimuli are shown in Figure 5.6. The pacman shape was recognized as familiar, as evidenced by a preference for the circle shape. Part 2 of the procedure was then administered and it simply followed the design of Experiment 1. The expectation is that if the category learning from Part 1 of the procedure, in particular, the representation of the invariant pacman shape, can interfere with the Gestalt-based perceptual

Familiarization Trials oo

Test Trials

Figure 5.6 Examples of the familiarization stimuli and test stimuli used in Quinn and Schyns (2003). If the infants can extract the invariant pacman from the familiar patterns, then they should prefer the circle shape over the pacman shape during the test trials.

organization that was observed in Experiment 1, then the preference for the pacman shape that was observed in Experiment 1 should no longer be observed. In fact, if the representation of the pacman shape carries over from Part 1 to Part 2 of the procedure, one would expect the opposite result, that is, the infants should continue to prefer the circle. The latter result is what was observed and it suggests that perceptual units formed during category learning can be (1) entered into a perceptual system's working "featural" vocabulary, and (2) available to subsequent object recognition processes. The bias set by the Gestalt principle of good continuation is thus soft-wired and subject to interference. More generally, an individual's history of categorization will affect their subsequent object parsing abilities. In the following section of the chapter, the issue of how categorization emerges as a core process during the period of early infancy will be considered more fully.


In addition to developing representations for coherent objects depicted in a single presentation of visual pattern information, members of the human species must at some point during development become capable of forming representations that are inclusive of numerous objects appearing over a

more extended period of time. In compiling these representations inclusive of multiple items, individuals are apparently detecting some basis of equivalence among them. This basis could be perceptual, functional, conceptual, or some even more abstract combination of the attributes possessed by the items. Because representations of this nature are normally developed for categories of items (e.g., dogs, chairs), I have referred to them in past writings as category representations (Quinn, 2002b). Category representations of many common kinds of objects may be essential to (1) organizing memory, and (2) permitting us to respond to many novel objects with familiarity. The latter occurs because of the recognition of equivalence among certain attributes detected from the objects, and maintained in their category representations (Murphy, 2002).

Categorization is considered to be a critical cognitive ability because a system of mental representation that lacked category representations would be dominated by unrelated instance information and would face the problem of having to respond anew to each novel object encountered (Smith & Medin, 1981). Indeed, the importance of category representations to daily cognitive functioning has led Thelen and Smith (1994) to argue that categorization is the "primitive in all behavior and mental functioning" (p. 143).

Although there has been a historical tradition among scholars of cognitive development to consider the ability to form category representations to be an achievement of childhood (Bruner, Olver, & Greenfield, 1966; Vygotsky, 1962), more modern work has focused on the abilities of infants and toddlers to respond categorically to common object types (Cohen & Strauss, 1979; Mandler & McDonough, 1993; Mervis, 1987; Oakes, Madole, & Cohen, 1991; Quinn & Eimas, 1996b; Waxman & Markow, 1995; Xu & Carey, 1996; Younger, 1990). The chapter will now consider the evidence on categorization by these younger participants, with emphasis on studies of categorization of realistic photographic exemplars of animals and artifacts conducted with 3- to 4-month-olds in the author's laboratory. Particular issues of current contention include exemplar versus prototype storage of category information in memory, the perceptual versus conceptual basis for early object categories, and the relative roles of learning occurring within the laboratory versus knowledge acquired prior to arrival at the laboratory.

Positive Thinking As The Key To Success

Positive Thinking As The Key To Success

Download this Guide and Discover How To Find And Monetize on Your Expertise And Strengths. Inside this special report, you'll discover: How positive thinking is one of the key factors in a successful life. Five ways and tools to help you stay positive. Use these to help you keep on track. Case studies that'll inspire you to stick to your dreams. Plus much, much more.

Get My Free Ebook

Post a comment