The Wiley-Blackwell Handbook of Childhood Cognitive Development

Usha Goswami

• Publisher: Wiley-Blackwell
• Year: 2011
• Print ISBN: 9781118586662, 9781405191166
• eBook ISBN: 9781444351736
• Edition: 2
• Topic: Psychologie
• Subtopic: Kognitive Psychologie & Kognition

This definitive volume is the result of collaboration by top scholars in the field of children's cognition.

• New edition offers an up-to-date overview of all the major areas of importance in the field, and includes new datafrom cognitive neuroscience and new chapters on social cognitive development and language
• Provides state-of-the-art summaries of current research by international specialists in different areas of cognitive development
• Spans aspects of cognitive development from infancy to the onset of adolescence
• Includes chapters on symbolic reasoning, pretend play, spatial development, abnormal cognitive development and current theoretical perspectives

PART I

Infancy

The Origins of Cognitive Development

The first section of the revised Handbook has been expanded to take account of recent developments in infancy research. One of the most remarkable findings from the last decade has been the statistical learning power of the infant brain. By processing sensory features of the input, and correlations and dependencies between these features, the infant brain basically learns about dynamic spatio-temporal structure, across modalities. Some relevant data are discussed in the chapters by Quinn; Baillargeon Li, Gertner, and Wu; Waxman and Leddon; and Bauer, Larkina, and Deocampo. Sensory statistical learning enables 2-month-old infants to learn visual transitional probabilities between abstract geometric shapes (Kirkham, Slemmer, & Johnson, 2002), and 3-month-old infants to learn to distinguish vehicles and animals on the basis of motion cues alone (when watching point light displays, Arterberry & Bornstein, 2001). Learning auditory conditional probabilities enables infants to extract structural properties from language input, for example the phonotactic patterns of language (the sounds that make up the language, and the orders in which they can be combined, see Tomasello, chapter 9), and the phonetic elements that comprise a particular language (e.g., Kuhl, 2004). Neural systems that can learn the patterns or regularities in environmental input captured by conditional probabilities can, in principle, acquire complex cognitive structures like language and concepts. On this kind of theoretical account, experience- dependent learning is the key to cognitive development, not the possession of innate “pre- knowledge.” Theoretical constructs such as a “Language Acquisition Device” for acquiring syntax (Chomsky, 1957) are not required to explain the complexity of infant learning.

Machine learning studies show that statistical learning algorithms other than conditional probabilities can be extremely powerful. For example, research in machine learning has also discovered algorithms enabling explanation -based learning. As discussed by Baillargeon et al. (chapter 1), in explanation-based learning a machine can generalize from a single example by explaining to itself why the training example is an instantiation of a concept that is being learned (DeJong, 2006). The machine uses background knowledge (prior domain knowledge) to constrain the inferences made. In their chapter, Baillargeon and her colleagues illustrate that physical reasoning by infants follows similar principles. They argue that infants form distinct “event categories” (occlusion, containment, support) about the physical world, apparently learning about each category separately. The data show that perceptual variables such as height that are identified in one category (e.g., by 4-month-olds for occlusion) are not necessarily generalized to another category (e.g., height is only identified as an important variable for covering by 12-month-olds). This appears surprising, as the height variable is equally relevant to both categories.

On Baillargeon’s data, infants appear to acquire event-specific expectations before event-general principles. However, this would be expected if a neural networks metaphor is applied. If the “physical reasoning system” of the infant is based on neural networks that are active in response to the sensory information experienced by the infant during each discrete event, then different types of “events” will be learned about separately. Other types of events will activate different neural networks, although there might be some overlap. Further, if infants are exposed to relevant discrete events purposely and incrementally within a category (as in the “teaching” experiments discussed by Baillargeon et al.), they will acquire event-specific knowledge earlier than if they lack such exposure. The experience-dependent construction of the neural networks responding to the spatiotemporal information in physical “events” such as covering will in itself yield this outcome.

The chapters by Meltzoff, Gergely, and Carpenter show that, when these perceptual learning mechanisms are supplemented with information gained by observing other agents and by acting oneself, then learning about the inner experiences or the mental life of others becomes possible. Meltzoff (chapter 2) argues that the first common code between self and other is the ability to map the actions of other people onto the actions of our own bodies. Action representation is the “supramodal” code enabling the infant to see others as “like me.” Meltzoff suggests that the infant’s experience with action productions and the consequences of actions enables a privileged understanding of people as distinct from other objects. A key developmental mechanism, available from birth, is imitation, which suggests an intrinsic link between action perception and production. Imitative ability depends on active intermodal mapping, the “supramodal code.” The child, even the newborn, can watch the movements of other people and recognize that “those acts are like these acts” or “that looks the way this feels.” Reciprocal imitation (e.g., copying games) enables further insight into intentionality. Meltzoff’s view is that seeing others as “like me” enables bidirectional learning effects and is central to understanding persons within a framework involving goals and intentions. Via their own copying actions, infants gain understanding of the acts of others. By watching others copy them, infants learn about themselves and the consequences of their own potential actions. Meltzoff argues that the result is a child who discovers facets of other minds through analogy with his or her own mind, and who simultaneously discovers powers and possibilities of the self through observing and imitating others. He also makes the critical point that effective teaching and learning is linked to one’s interpretation of the motivations and goals of others.

In chapter 3, on the early understanding of kinds of intentional agents, Gergely presents different evidence relevant to the same problem of knowing other minds. He suggests that being sensitive to information about what another “knows” on the basis of perceptual information about their line of regard is a basic and innate adaptation that has evolved in other social species as well (such as apes, crows, and scrub jays). For example, a bird may re-cache its food store having observed another bird see it. With respect to actually representing the content of the mind states of another, which appears to be specific to the human species, Gergely argues for two kinds of adaptations. One supports the recognition and interpretation of intentional agents performing goal-directed actions, and the second supports the recognition and interpretation of communicative agents performing acts of informational transfer (natural pedagogy). Gergely suggests that the developmental mechanisms enabling these adaptations include the ability to detect causal patterns of distal contingent reactivity between agents, such as taking turns; infants’ preference for direct eye gaze; their sensitivity to ostensive cuing (e.g., eyebrow raising); and their sensitivity to “motherese” (the exaggerated prosodic register that we use to speak to babies). Intentional agents are those that babies perceive to make rational choices of action between multiple alternatives. Communicative agents are those that babies perceive as offering turn-taking patterns of communicative contingent reactivity. Gergely also argues that the system for detecting, interpreting, and learning about intentional agents may be developmentally independent of the separate system for mindreading, and that part of cognitive development might be the integration of these two systems.

Carpenter (chapter 4) addresses similar developmental issues from the complementary perspective that babies are motivated to share psychological states like attention and goals with others, and to do things the way that others do. This “shared intentionality” is what enables humans (and only humans) to engage in collaborative activities and to create cultural practices and institutions. Like Gergely, Carpenter invokes evidence from non-human species to show that some of infants’ social cognition skills are not species specific. Both apes and babies can understand the goals and intentions of others, can understand others’ perceptions, and can understand their knowledge or ignorance. The difference is that only babies participate in activities that require shared intentionality. The motivation to engage in joint attention, for example, is strong – infants will turn away from an engaging new event in order to draw their mother’ s attention to it so that they can attend to it together. Infant and adult know together that they are sharing attention and attitudes. Infants also keep track of the knowledge and experiences that they have shared with others in the past, what “we” know together, and comprehend and produce communicative gestures with this in mind. This enables joint collaborative activity, which is not found in the animal kingdom. Carpenter’s thesis is that shared intentionality – the motivation to share psychological states and experiences – is unique to humans.

Quinn (chapter 5) considers the advantages of having a mind–brain system that categorizes experience. He argues that categorization is fundamental to cognition, enabling efficient learning and memory, and reducing the complexity of the external world. He considers evidence relevant to how infants go about the task of dividing the unlabeled world into like entities, which are then stored mentally as category representations. He explains Rosch’s (1978) premise that the perceptual world has inbuilt structure in the form of bundles of statistically correlated attributes, to which the infant brain is sensitive. Quinn considers whether initial attention to correlation is at the “basic” categorical level of dogs and cars, or at the “superordinate” categorical level of animals and vehicles. He concludes that global categories emerge before basic categories because of cumulative instance-based learning. When a small number of exemplars has been experienced, the representation will be fairly general (global, for example that self-propelling animates with four legs are animals). When a large number of exemplars has been experienced, the representation can be more detailed (for example, enabling medium-sized four-legged animates to be divided into cats versus dogs). The key learning principle is exemplar-based and cumulative, which means that the number of exemplars experienced will produce changes in the inclusiveness of the category representations (global versus basic) and apparently qualitative changes in representation (from perceptual to conceptual). The similarities with Baillargeon’s enrichment account and the developmental processes suggested by the neural networks metaphor are striking.

Bauer et al. (chapter 6) discuss the early development of declarative and non-declarative memory. They argue that this distinction is vital for developmental scientists, as the two types of memory rely on different neural substrates and develop in different ways. Non-declarative or implicit memory (such as learning habits and skills) involves non-conscious abilities and depends on incremental learning. Declarative or explicit memory involves what we usually mean by “remembering” – the conscious recall of places, dates, events, and so on. Declarative memory is fast (it can be formed on the basis of a single experience), can be fallible (because memory traces degrade and retrieval failures occur), and is flexible. Bauer et al. show that the development of new paradigms such as infant habituation and deferred imitation has enabled the documentation of really quite remarkable mnemonic capacities in infancy. They focus on event memory, as our memories of events define the self – who we are is who we were and what we did. Event memory becomes increasingly autobiographical across the early years of life, and experiments suggest that consolidation and storage processes, rather than retrieval processes, are the major sources of developmental change. An event that is not well encoded or stored cannot be retrieved. Bauer et al. also provide an overview of the neural structures supporting memory, as available neuroscience data are quite extensive here.

Waxman and Leddon (chapter 7) focus on early word learning and its critical impact on cognitive development. As infants learn words, they learn symbols that stand for or refer to something in the real world, and therefore they learn a code that they can use to manipulate their knowledge about the real world – a symbolic system. Waxman and Leddon describe the close interaction between early conceptual development and early linguistic development in detail. They argue that infants are born equipped with an innate expectation that words will refer to commonalities among objects. These commonalities can be of many kinds, for example taxonomic (dogs, cats), functional (pulls, cuts), thematic (bread goes with butter), or property-based (has wings, is red). As in Baillargeon’s account, an initial core expectation is assumed to be fine-tuned via experience. As in the social cognition chapters, there is also a causal role for infants’ innate propensity to attend to and interact with other people. Infants show a special interest in “people sounds” – the sounds of language. They rapidly become perceptually tuned to the phonologic, prosodic, and morphologic elements characterizing their native language, and, very soon, novel words guide attention to objects and highlight commonalities and differences between them. The ability to produce words oneself (naming) has powerful cognitive consequences, promoting the formation of object categories, and a means of tracing the identity of individual entities within these categories. Waxman and Leddon argue that words are powerful engines for conceptual development.

References

Arterberry, M. E., & Bornstein, M. H. (2001). Three-month-old infants’ categorization of animals and vehicles based on static and dynamic attributes. Journal of Experimental Child Psychology, 80,333–346.

Chomsky, N. (1957). Syntactic structures. The Hague/Paris: Mouton.

DeJong, G. (2006). Toward robust real-world inference: A new perspective on explanation-based learning. ECML06, The Seventeenth European Conference on Machine Learning (pp. 102–113).

Kirkham, N. Z., Slemmer, J. A., & Johnson, S. P. (2002). Visual statistical learning in infancy: Evidence for a domain general learning mechanism. Cognition, 83, B35–B42.

Kuhl, P. K. (2004). Early language acquisition: Cracking the speech code. Nature Reviews Neuroscience, 5, 831–843.

Rosch, E. (1978). Principles of categorization. In E. Rosch & B. B. Lloyd (Eds.), Cognition and Categorization (pp. 27–48). Hillsdale, NJ: Erlbaum.

CHAPTER ONE

How Do Infants Reason About Physical Events?

Renée Baillargeon, Jie Li, Yael Gertner, and Di Wu

Introduction

As adults, we possess a great deal of knowledge about the physical world: for example, we realize that an object continues to exist when hidden, that a wide object can fit inside a wide but not a narrow container, and that an object typically falls when released in midair. Piaget (1952, 1954) was the first researcher to systematically investigate the development of infants’ physical knowledge. He examined infants’ responses in various action tasks and concluded that young infants understand very little about physical events. For example, after observing that infants younger than 8 months do not search for objects they have watched being hidden, Piaget proposed that young infants lack a concept of object permanence and do not yet understand that objects continue to exist when hidden.

For the next several decades, Piaget’s (1952, 1954) conclusion that young infants possess little or no knowledge about the physical world was generally accepted. (For reviews of this early research, see Bremner, 1985; Gratch, 1976; Harris, 1987). This state of affairs began to change in the 1980s, however, when researchers became concerned that exclusive reliance on action tasks as an investigative tool might underestimate young infants’ physical knowledge. In order to search for an object hidden under a cloth, for example, infants must not only represent the existence and location of the object, but they must also plan and execute the appropriate means–end actions to retrieve it. Thus, young infants might represent the object but still fail to search for it because (a) they are unable to plan or execute the actions necessary to retrieve it (e.g., Baillargeon, Graber, DeVos, & Black, 1990; Diamond, 1991; Willatts, 1997), or (b) they can plan and execute these actions but lack sufficient information-processing resources to simultaneously represent the hidden object and carry out the actions required to retrieve it (e.g., Hespos & Baillargeon, 2008; Keen & Berthier, 2004; Lockman, 1984; see also Munakata, McClelland, Johnson, & Siegler, 1997; Shinskey, 2002; Shinskey & Munakata, 2001).

The preparation of this chapter was supported by a grant from the National Institute of Child Health and Human Development to Renée Baillargeon (HD-21104). We would like to thank Jerry DeJong and Cindy Fisher for many helpful discussions.

These methodological concerns led investigators to seek alternative approaches for exploring young infants’ physical knowledge. Their research efforts can be roughly organized into three successive, overlapping waves. The first wave established that, contrary to Piaget’s (1952, 1954) claims, even young infants possess some expectations about physical events. The second wave began to systematically examine the development of infants’ physical knowledge and brought to light striking patterns of successes and failures in infants’ responses to physical events. Finally, the third, ongoing, wave builds on these preceding efforts and attempts to specify both how infants reason about physical events and what cognitive architecture makes this reasoning possible. In what follows, we first briefly review findings from the first and second waves. In the remainder of the chapter, we focus on the third wave and present a three-system account of how infants reason about physical events.

First Wave: The Competent Infant

One of the major alternative approaches used to explore young infants’ physical knowledge relies on the long-established finding that infants (like older children and adults) tend to look longer at stimuli they perceive to be novel as opposed to familiar (e.g., Fantz, 1 956). Looking-time tasks have two main advanta...

Table of contents

• Cover
• Series Page
• Title page
• Copyright
• Dedication
• Acknowledgements
• List of Contributors
• Introduction
• PART I: Infancy
• PART II: Cognitive Development in Early Childhood
• PART III: Topics in Cognitive Development in Childhood
• PART IV: Theories of Cognitive Development
• Index