Exploring Science: The Cognition and Development of Discovery Processes

By David Klahr; Foreword by Herbert A. Simon)
MIT, A Bradford Book, Cambridge, MA, 2002
255 pp. 55. illus. b/w. Trade, $37.95/£25.50; paper, $22.00/£14.50
ISBN 0-262-11248-5; ISBN: ISBN 0-262-61176-7

Reviewed by Amy Ione
The Diatrope Institute
PO Box 12748
Berkeley, CA 94712-3748 USA.

ione@diatrope.com

Bombarded with images of war, I turned to David Klahr's Exploring Science: The Cognition and Development of the Discovery Process hoping it might provide a welcome respite from the CNN effect. While unable to change world dynamics, it did provide an opportunity to think about all that creative insight adds to the quality of life. The book's strength derives from the way it amplifies Herbert Simon's idea that scientific discovery is neither impenetrable nor ineffable. Less successful is the examination of the questions of what scientific thinking is, how it can be studied, and the nature of its developmental course. Still, to its credit, Exploring Science acknowledges the degree to which each solved scientific problem raises new problems for study and that scientific advance is very much a social activity. Yet in my opinion, although this well-written account accentuates the collaborative nature of problem solving, the problems solved in designing this theory made the results strikingly removed from the broader framework of human living, especially in light of all of the havoc now taking place in our world. This lacuna highlights the range of problems even the most creative scientists cannot solve, although this is not to suggest it was a 'bad' book, or a book that failed to present convincing experiments. Rather, the narrow context of the studies, the stated aim to show that science is a highly social activity, and the goal of expanding our view of cognition and the discovery process so as to emphasize the ways scientific processing is similar to normal cognitive processing combined to leave the impression that the conclusions were more abstract than convincing. While the nine laboratory experiments presented, to be sure, demonstrate that we can equate scientific discovery with common activities, the experiments hardly compare to the complex world we inhabit, where the carefully crafted laboratory situation does not apply.

I do not mean to imply that Klahr failed to affirm the work's limitations. Indeed he points out that the projects discussed only provide partial answers to the psychology of scientific reasoning and that the purpose of the documented studies was to in fact investigate scientific discovery processes in the psychology laboratory. This was done by the researchers efforts to create discovery contexts that evoke the kind of thinking characteristic of scientific discovery in the "real world." Throughout categories are well defined, clarification of how theoretical and experimental models develop are useful, and the work succeeds in integrating two disparate approaches: the content-based and the process-based. Klahr also explains his attempts to broaden a field in which historical psychological investigations have tended to produce diametrically opposing views of the development of scientific reasoning skills. This book's goal of synthesis achieves more overall balance. The contributors also use their experimental work to flesh out the conflicting claims (theoretical and empirical) emerging from the "scientist as child" debate as well as debates centered around ideas about how children and scientists really think. In engaging with all of these elements, problem solving is emphasized and defined as the quality of scientific reasoning that is key to scientific discovery. Characterizing the work in terms of the scientific discovery process that was pioneered by Newell and Simon (1972), which defines problem solving as a search for a path that links an initial state to a goal state works well. The book also relies heavily on Marvin Minsky's "frame" notation when discussing their SDDS (scientific discovery as dual search) model, one that is intended to depict the goal structure that is generated in a broad range of scientific reasoning.

As is often the case when the topic is the discovery process, omitted research was of greatest concern. While successful in looking beyond the limitations of historical approaches, the analysis did not quite feel in touch with contemporary cognitive studies. Little effort was made to incorporate neurological data comparing learning in adults and children, brain processing studies on task domain analysis, etc. Seeking to formulate theories of different aspects of the scientific process in the form of a computational model is no doubt served by the task driven approach to some degree. Yet, given all that we are learning about brain processing now through direct studies of the brain processing, it seemed that depending on psychological studies of participants' thinking processes in contexts designed to present several important characteristics of scientific problems was too limited. The theory too quickly assumes that constructing a computational model of intelligence need not factor in data that directly relates to the cognitive process. Perhaps neurological studies have nothing to add, but I believe some mention of why those involved in the research reached this conclusion would make for a more robust, more convincing argument.

Scientific discovery is also inadequately defined, in my view. I would have liked to see more mention of how anomalous cases fit into the SDDS model. For example, after Wilhelm Conrad Röntgen announced his discovery of the X-ray in 1895, others acknowledged that they had perceived various anomalies that, had they investigated, might have led them to be the first to realize how the use of invisible radiation can produce and record a visible image of an invisible object property. Klahr and his colleagues do not account for why one person might ignore information that provokes another to look more closely. Also, although the chapters do discuss prior knowledge and competently compare the approaches of children and those of adults, they fail to comprehensively address the degree to which a prior knowledge base influences a conclusion. In addition, the laboratory studies as outlined fail to address how lay people and scientists incorporate preconceptions into their analysis and why this often leads them in orthogonal directions. To continue with Röntgen's case, while Roentgen was enthusiastic about his revolutionary discovery, his wife had a radically different response. After her husband introduced her to it by showing her an x-ray image of her (now skeletal appearing) hand, Frau Roentgen was convinced it was an omen of death and never returned to his laboratory. More generally, preconceptions often halt the discovery process. For example, despite the ongoing questions raised by scientists and society at large, one of the reasons that the Ptolemaic, earth-centered solar system lasted for as long as it did was that it appeared to match the world of our experience better than the sun-centered view. People could see that the sun appeared to circle the earth and, over time, the very longevity of the theory supporting this perception was one rubric used as evidence of the earth-centered theory's validity when alternatives were proposed.

Leonardo readers will no doubt find the failure to mention art disappointing. The analysis cried out for it. Many of the processes analyzed are a part of artistic discovery as well as discovery in general. Moreover, it is in the 'discovery' domain that art and science are most easily intertwined. Ironically, it seems the researches were searching for a means to speak about what is popularly termed the art of science and failed to make the connection explicitly, for whatever reason. Nonetheless, in their task domain analysis they do analyze the ways in which those engaged with a problem "play" with factors to "feel" their way to a "correct" solution.

All of the above leads me to conclude the results of this research are more effective in raising the question of what we mean by scientific discovery than answering it. I was disappointed that the chapters did not examine why we hold so many views on cognition and development in the scientific discovery process, and did not do so in a way that reviewed alternative theories about creative discovery (proposed by people such as Amabile, Csikszentmihalyi, Gardner, Simonton, Sternberg, Weisberg, etc.). Robert J. Sternberg's anthology on creativity (The Nature of Creativity, 1988), for example, concludes with a chart that maps twenty-one cognitive characteristics of creative people against sixteen researchers in this area who contributed to his book. What is most striking about the chart is the lack of agreement among those who write on this subject. Although Sternberg's chart is not specific to scientific creativity, it does show that there are diverse views on cognitive characteristics of creative people. Klahr, who incorporates the work of many psychologists who have proposed theories in this area (e.g., Piaget, Bruner, Gruber, Perkins, Simon, etc.), is more concerned with demonstrating that the SDDS model, although not complete, is correct than he is in placing its weaknesses in dialogue with the views of others. The very fact that this book claims that the weak methods invoked by scientists as they ply their trade are the same ones that underlie all human cognition makes me wonder if this publication would have seemed more convincing if some of the competing theories had been more fully integrated into this discussion.

In summary, although Klahr and his collaborators never suggest that the average person could walk into a lab and do what a trained scientist might do, their equation of scientific discovery with common cognition nonetheless fails to cohere. It is not that the argument per se is unconvincing. The writing does successfully convey ways in which participants who attempt to solve the problems posed by discovery tasks invoke many of the same higher-order cognitive processes that scientists use as they practice their trade. Still the experiments and the limited literature review constrain the analysis.

Bibliography:

Newell, Allen, and Herbert A. Simon. Human Problem Solving. Englewood Cliffs, NJ: Prentice-Hall, 1972.

Sternberg, Robert J., ed. The Nature of Creativity: Contemporary Psychological Perspectives. Cambridge: Cambridge University Press, 1988.