a concert on the violin while learning to play the instru- ment. Butler's insight-that there are no distinct training and performance modes in development-succinctly.
A Dynamic Systems Approach to the Development of Cognition and Action Esther Thelen and Lhda B. Smith Cambridge, MA: The MIT Press, 1994 Hardbound, 376 pages, $50.00. ISBN 0-262-70059-x Reviewed by David J. Lewkowicz and Robert Lickliter
life, said the English novelist Samuel Butler, is like giving a concert on the violin while learning to play the instrument. Butler’s insight-that there are no distinct training and performance modes in development-succinctly c;iptures the activity-dependent nature of the developmental process. For example, imagine a baby learning to walk. Although babies spend a great deal of time falling down, getting up, and falling down again, as far as some parents are concerned, their child’s walking seems to magically appear de nouo one day. Other parents, however, d o appreciate the importance of their baby’s ups and downs and believe the old saying that “practice makes perfect” in the emergence of walking. Butler’s insight, however, suggests a third perspective on the emergence of such novel behavior: namely, that there are no practice modes in development-all activity counts in selecting, constructing, and maintaining behavioral outcomes. This fundamental role that self-generated activity plays in the development of all behavioral functions, ranging from motor behavior to perception, cognition, and social behavior, is a central theme of the dynamic systems approach that is masterfully explicated in the recent book by Esther Thelen and Linda B. Smith. In this book the authors (both developmental psychologists) apply the principles of nonlinear dynamic systems t o motor, perceptual, and cognitive development and the result is a remarkable synthesis and advancement of developmental science. Indeed, it would not be an exaggccition to say that this volume offers a revolutionary perspective on the development of cognition and action. According to Thelen and Smith the primary thrust of dcvclopment is the generation of novel structure and behavior. In their attempts to explain this complex process, Thclen and Smith eschew the classic dualisms of structure vs. function, nature vs. nurture, brain vs. behavior, perception vs. cognition, mind vs. body, competence vs. performance, and learning vs. development. Their rejection of dualistic thinking results in a radical departure from current cognitive theory and provides a view of development in which stability and change are accommodated under a single theoretical umbrella. This umbrella, o f course, is dynamic systems theory, an approach that is doubtlessly familiar to many of the readers of this journal. What makes this book different from other books on the topic of dynamic systems, and more useful than many in this genre, is both its breadth and its explicit focus on the developmental processes underly-
ing the emergence of motor, perceptual, and cognitive functions. Indeed, we believe Thelen and Smith’s book is unique in this area in that it provides multiple examples of the successful application of the dynamic systems approach to the study of human development as well as a methodological agenda for the generation of new research programs. Thelen and Smith begin their book by discussing the shortcomings of many of the major theoretical systems associated with the study of human development (e.g., maturationist, neurological, rationalist-nativist, and information processing) and note that all of these approaches have a common teleological core that implicitly presumes an “end-state”before the developmental process even begins. The authors then go on to reject simplc. cause-and-effect models that rely on predeterminism. linearity, and reductionism. In their place, they offer ;I comprehensive approach to the development of behavior that focuses on the processes underlying change in dynamic, self-organizing systems. These systems have a history, can create novelty out of contextual and organismic conditions, and do not have an ”end-state”cocletl anywhere. The early chapters of the book are organized to showcase the authors’ own research programs on early motor (Thelen) and cognitive (Smith) development to illustrate the explanatory power of this approach for complex behavioral phenomena. In these chapters the authors also utilize the power of their approach to show how previous interpretations of findings in these research areas are misguided in relying on concepts o f innate programs or innate knowledge structures. In subsequent chapters Thelen and Smith ;iddress the relation between morphology, neural processes, and behavior in promoting developmental change. In relating these three, the authors rely heavily on N. A. Bernstein’s (1967) concept of indeterminacy and Gerald Edelman’s (1987) concepts of selection and reentry (the interrelating of several or many simultaneous perceptual and motor representations) to provide the means by which novel neural and behavioral structures and functions may be “soft-assembled”in real and developmental time. Of particular interest to the readers of this journal is the authors’ argument that the dynamics of behavioral phenomena and the known dynamics of neural phenomena should be consistent with one another. Most importantly. Thelen and Smith argue that a reductionistic approach that resorts to interpretations based solely on neural factors leads to infinite regress in our explanations of the process of the emergence of novel structure and behavior. In their dynamic systems approach, the neural aspects of human functioning are part-and-parcel of the individual’s entire developmental system. N o single element or level in this system necessarily has causal primacy or privilege. Indeed, Thelen and Smith’s approach to development virtually requires that equal status be given to all contributing factors at all levels of analysis. A central question for students of perceptual and
cognitive development is how perceptual categories, which are the foundation of human cognition and action, emerge during development. What are the processes and mechanisms that allow meaning to be assigned to the multiple and continuous sensations of everyday life? ’I‘helen and Smith devote a major section of the book t o a discussion of this question and propose that the answer must lie in (1) the recognition of the fundamental unity of perception and action and (2) the critical importance of the temporal association of multimodal information in creating the link between the mind and the world. In their active exploration of the world, infants receive massive amounts of time-locked,multimodal sensations. The authors argue that infants discover action and object categories through the cross-correlation o f these multimodal experiences. In their own words, ”what infants know and how they act are selected continuously and dynamically from what they encounter and how they act. Sensory integration is the primitive, not the derived state, and knowledge is limited not primarily by deficits in storage but in the ability to adequately sample and thus categorize the world” (p. 21 1 ). This position obviously challenges much contemporriry work in the field of cognitive development and will likely engender controversy among students of cognition. In the remaining chapters of the book Thelen and Smith discuss and explain developmental phenomena as diverse as memory, traversing slopes, word knowledge, reaching, and the A-not-B error within the dynamic systems framework.For example, Thelen and Smith demonstrate the value of dynamic systems principles in explaining the development of reaching in human infants.They show through the detailed longitudinal study of individual infants’ performance how different babies arrive at very different solutions to the common problem of retrieving an object. Through detailed examination of the day-today variations in arm movements, the authors convincingly demonstrate the unique developmental pathways that different infants utilize in their eventual solution to the problem of reaching. By doing so, Thelen and Smith elucidate the distinctive, contextdependent, and self-organizednature of development. In their own words, the “solutionswere discovered in relation to their own situations, carved out of their individual landscapes, and not prefigured by a synergy known ahead by the brain or the genes” (p. 260). Thelen and Smith conclude the book with a considered look at the “hard problems” facing contemporary developmental science. In their opinion, one of the key unresolved areas is the traditionally thorny issue of value and motivation. In their discussion they resurrect the important but often overlooked contributions of Kurt Lewin, who recognized the significance of a systems approach to behavior over half a century ago. In particular, Lewin focused on the central role of value and motivation in directing developmental change. Similarly,
the concept of value is also very important to Thelen and Smith’s dynamic systems theory because it provides a primary mechanism by which perceptual categorization is achieved. Value provides the motivational impetus for the organism to seek certain forms of stimulation and not others. Thus, very early in development, infants might be biased to seek contact or warmth, prefer light over dark, seek objects that can be sucked or touched, etc. Once objects or events that are preferred are encountered, different properties of these objects or events can then be differentiated through recurrent experience. According to Thelen and Smith, infants come into the world with a set of values or hedonic valences and the authors suggest that such values are somehow epigenetically acquired. What the authors do not offer, however, is a mechanism by which epigenetic forces confer positive or negative valence on particular forms of stimulation. One such possible mechanism was proposed decades ago by the comparative psychologist T. C. Schneirla (1959). After reviewing the behavioral responsiveness and organization of young animals across different phyletic levels, Schneirla noted that the earliest forms of behavior in developing organisms consist of approach and withdrawal reactions. He proposed that the principal factor that accounted for these approach/withdrawal reactions was the effective intensity of stimulation (i.e., the overall level of stimulation engendered by factors internal and external to the organism). He noted that when the effective intensity of stimulation is low, organisms typically approach the source of stimulation, whereas when the effective intensity of stimulation is high, organisms typically withdraw. Very important to this notion of approach/withdrawal is that it is a noncognitive construct that does not imply prior knowledge of stimuli or their potential consequences. What Schneirla’s theory offers to Thelen and Smith’s dynamic systems approach is a metric for predicting levels of sensory stimulation that will either have positive or negative valence based on the specific level of effective intensity. In other words, we believe approach/withdrawal theory offers a means to operationalize value and its role in the developmental origins of initial perceptual categorizations. As this necessarily brief review attests, the range of issues and phenomena covered in this extraordinary book is extensive. At the same time, a striking feature of this book is its thematic cohesion and integration across diverse disciplines and levels of analysis. The authors have obviously labored long and hard to make it possible for the reader to critically consider complex and abstract issues. They also have provided a companion volume of edited chapters (Smith & Thelen, A Dynamic Systems ilpproacb to Development: Applications, MIT Press, 1993) in which these issues are considered in the context of specific research programs. The publisher (MIT Press) of this volume, as well as the companion volume, Book Reviews
51.3
has done a laudable job in producing an attractive book of high quality. In sum, perhaps the highest praise we can give Thelen and Smith’s most recent book is to confess that we both found ourselves being transformed intellectually in the process of reading it. Nearly every page required us to reflect on, reconsider, and in some cases reformulate our varied assumptions and viewpoints about the process of development. We sincerely hope that this book receives the wide rcadership that it deserves and in doing so helps to realize the revolution in developmental science it argues for. Reprint requests should be sent to either David J. Lewkowicz, NYS Institute for Basic Research, 1050 Forest Hill Rd., Staten Island, NY 103 14 o r Robert Lickliter,Department of Psychology, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061-0436.
REFERENCES Bernstein, N. (1967). Coordination and regulation of movements. New York: Pergamon Press. Edelman. G . M . (1987). Neural Darwinism. New York: Basic l3ooks. Schneirla, -1’. C. (1959). An evolutionary and developmental theory o f biphasic processes underlying approach and withdrawal. In Jones, M . R. (ed.) Nebraska symposium on motivation. Lincoln: University of Nebraska Press, pp. 1-42.
The Cognitive Neurosciences Edited by M. S. Gazzaniga Cambridge, MA: The MIT Press, 1994 Hardbound, 1447 pages, $95.00. ISBN 0-262-07157-6 Reviewed by James M. Sprague
This handsome volume is the latest of several excellent compendia in the field of cognitive neuroscience, and is the largest (1447 pages) and most comprehensive of the group. It consists of 92 chapters, ranging in subject matter from molecular and cellular plasticity, neural and psychological development, sensory systems, motor systems, attention, memory, language, thought and imagery, emotion, evolutionary perspectives, and consciousness. These general subjects are organized by a distinguished group of section editors and the whole is presented under the Editor-in-Chief,Michael Gazzaniga.The book is I>eautifiillyproduced and published by the Press of the Massachusetts Institute o f Technology. The chapters represent in summary form a vast amount of information, experimentation, and critical evaluation. Some very interesting hypotheses, which are bound to stimulate new lines of research and new thinking, are included. Because of the constraints of length of
text, the literature discussed and referenced is chiefly that of research done over the past several decades, which does not allow adequate consideration of classical background material on which present concepts are based. Cognition is the most complex function of the brain, and in a hierarchical concept of the organization of the brain, it is arguably the “highest.” Such a function requires a broad and very comprehensive information base to understand its representation in the brain and the neural mechanisms that generate it. In short, cognition is a tough nut to crack. The aim of this volume is obviously to survey the available information in a form useful to nonspecialists as well as specialists, to students as well as experts. Cognition is not a single entity but is an emergent function of the neural integration of many systems and subsystems, both feedforward and feedbackward, afferent and efferent, which have been dissected and studied piecemeal in these chapters. All o f the contributions in this volume are important to achieve understanding, and many of them are elegant in their design and synthesis. The relation of their presentations to cognition are clearer and more compelling in some chapters than others. This is due partly to the material covered, and partly to the degree of boldness and imagination of some writers versus the conservatism of others. I n the effort to grasp the neuroscience basis of such a tenuous, yet vital function as cognition, the words of the Editorin-Chief in the Preface are trenchant-‘“Cognitive neuroscience walks a thin line. It must be exacting and build its foundation on the best and most stringent of observations about the mysteries of nature. On the other hand, it has to explore, in an intelligent and probing antl verifiable way, how primary data speak to the issues of how brain enables mind.” In a volume of this size and breadth of coverage, many readers will inevitably pick out for special attention the subjects most familiar to them. Yet in the eyes of this reviewer, the editor’s aim is to present the material in such a way that the reader will be stimulated to read (albeit over time) the whole volume from beginning to end. A Sdhtdry effort has been made by the writers to reduce the jargon familiar to specialists but a stumbling block to others, and when used they have usually defined it and explained its meaning. Thus, the nonspecialist and graduate student can read most of the chapters profitably. There is an extensive index and thc illustrations include a series of color plates so important in reading data derived from iniaging techniques. The chapters are not written for submission to refereed journals, but rather to inform and educate, and hence any attempt to “review” them for content antl point-of-view is inappropriate. The important function here is to present something of the direction and flavor of the research covered and its relation to cognition. I will attempt to fulfill that aim by picking out summaries.
