Quantification, Cognitive Psychology, and the Computer

Psychology became increasingly' quantitative in its orientation. The use of statistical methods for the evaluation of empirical results by' virtually everyone in the field hardly needs mention. Quite aside from the mental testing movement and the ubiquity' of statistics, there was a strong and growing tendency to construct theories having the precision of mathematical formulations. Major impetus was given to this development by the invention of information theory by communications engineers like Claude E. Shannon and mathematicians like Norbert Wiener; information theory, or concepts akin to those in the theory, were applied to a wide variety' of problems in sensation, perception, and learning by people like George A. Miller and Fred Attneave. Sensation, already quantitatively approached by investigators a century earlier, became again a major focus of mathematical thought in psychology with the development of sensory scaling by S. S. Stevens, W. R. Garner, J. G. Beebe-Center, and others. There was also much interest in mathematical learning theory, beginning perhaps with Clark Hull (although the trend can be traced back at least to Herbart) and the construction of mathematical models of learning by' Robert Bush and Frederick Mosteller,

William Estes, Frank Restle, and many others. Decision theory was applied to human problem solving and troubleshooting by individuals such as Ward Edwards. In social psychology a variety of balance models were proposed by people such as Helen Peak, Fritz Heider, and Theodore Newcomb. One version of balance theory holds, for example, that if A and B like each other and A and C dislike each other, then the structure of the trio A, B, and C will be relatively stable if B and C dislike each other but unstable if they do not.

The technical breakthroughs in computer technology had a profound impact on psychology'. Statistical analyses that used to take an inordinate amount of time to do by hand could be performed very rapidly, and complex analytical calculations that were too laborious to be practically feasible fell easily within the capability of even relatively modest high-speed computers. As a consequence, the degree of sophistication in the analysis of the data from typical empirical studies was substantially enhanced. Factor analyses, analyses of variance and covariance, parameter estimations, autocorrelation functions, curve fitting, power analyses, and other advanced statistical procedures that used to take days or weeks on a desk calculator could be performed routinely, efficiently and economically by computer in minutes or seconds on laptops and even on hand-held devices.

But more profound than the increased efficiency of statistical calculations on research data as such was the impact that the computer had on the psychology of thinking and problem solving in particular and of information processing in general. It is probably no exaggeration to claim that the modern computer generated the entire new field of cognitive science, including both what is called artificial intelligence (Al; computer programs that are, typically, superior to any real human expert in speed and accuracy in diagnosing disease within a circumscribed domain, flying an aircraft, finding water or oil, or predicting the weather) and the computer simulation of higher mental processes. Little progress occurred in cognitive psychology (indeed, to repeat, the words cognitive and cognition were generally viewed with derision by behavioristically oriented psychologists) during the decades immediately following the Gestalt work on insight by such workers as Kohler and Wertheimer. But by the late 1950s and early 1960s Noam Chomsky had proposed his transformational grammar, and Herbert A. Simon and his colleagues had invented computer programs that could solve problems in propositional logic; elaborations on these programs, which soon not only successfully simulated the steps that college students take in solving logic problems but actually also generated some of the great discoveries of 18th- and 19th-century natural science, won Simon a Nobel Prize in 1978.This work became the foundation for the huge and popular field of cognitive psychology.

Among many landmarks in cognitive psychology were Ulric Neisser s mid-1960s book of that title, Cognitive Psychology, which helped to establish the field; the sophisticated work of Walter Kintsch and his associates on the processing of text materials; and the attempts of Janies Greeno and his colleagues to simulate the learning of geometric proofs with genuine understanding. A major strategy in the field was an effort to describe precisely the mental representation of a problem and of the knowledge domain relevant to it, including how this representation is organized. The hope appeared to be that theoretical work that is as concrete and rigorous as that required to write such computer programs whose outputs are superior to, or are indistinguishable from, the behavioral output of living, learning, and thinking human research participants, would lead to a better understanding of such complex and fascinating psychological events in people.

The strictly precise quality of computer models of psychological processes fitted well with the earlier behaviorist emphasis on objectivity. Computer programs as heuristic representations of cognitive processes have typically been consistent with the mechanistic orientation of behaviorism, tempered by operationism. They can be viewed as a modern continuation of early trends of scientific materialism, quantification, and usually atomism and associationism as well.