Theory of Human Problem Solving

In the preceding chapter a taxonomy of pr oblem types was proposed, and some hypotheses set forth about the general kin ds of skills and knowledge needed to solve problems of each type.. The p resent chapter, adopting the same information-processing point of view as the previous one, undertakes three tasks. The first part of the chapter sets forth th e general theory of human problem Solving that has emerged from research in the past two decades, especially research that has employed the methods o f computer simulation and analysis of thinking-aloud protocols. The second part examines recent and ongoing research aimed at giving an account of the role of pe rceptual processes in problem solving and a description of the processes for gen erating problem representations, and research aimed at extending the theory to problem solving in domains that are rich in semantic information and less well structured man those that have been examined in the past. The third part of the chapter discusses some of the methodological issues that must be faced in using the methodologies of simulation and protocol analysis and, in general, to test detailed processing models of human cognitive performance. Since this chapter and the preceding one both discuss their topics in information-processing terms and since bo th review essentially the same body of evidence, they reach conclusions that are Irrgely complementary. The present chapter does take, however, a somewhat m ore sanguine view than the preceding one as to how far we have already progres sed in building and testing a coherent theory of human problem solving. The th eory discussed in this chapter applies both to Grccno's category ofproblems o f transformatiort and his category of problems of arrangement. The relation o f the theory to problems of inducing structure is discussed ;n Sur.an ;m«J Lea (1 974). For the purposes of the present