Cognitive Analysis, Computer Models, and the Wais

Fulkerson and Cartwright (1984) have an interesting approach to how the human brain analyzes Similarity-type tasks. Like other models, theirs makes certain assumptions about how the brain works, assumptions which are rarely clearly articulated. One of these assumptions is the idea that che brain processes information in a linear, orderly, and incremental manner. In fact, evidence is growing that the brain processing of higher cognitive material involves multiple simultaneous, often non-linear processes (which may influence one another) as well as linear and overlearned automatic strategies which may or may not be applicable to the problem at hand (from the computer's point of view). Such processes reflect a complex interaction between brain organization and development with life experience. From this perspective, models as generally proposed by psychologisrs oversimplify brain processes and miss essential aspects of brain processing. Until such models acknowledge and analyze these processes, they are unlikely to be very useful in predicting a substantial pan of the variance in even simple human behavior. Fulkerson and Cartwright attempted to study their model experimentally. Although the study can be criticized on a number of general methodological issues (e.g., the population). the study still illustrates the limits of the advocated theoretical approach. The authors need, for example, to combine actual and potential overlaps although the role of potential overlaps in the model is quite questionable. While total overlap correlates with the Similarities-score rota], such correlates are modest, accounting for some 16% of the total variance in three items and none in the fourth. Abstraction is a poor predictor for two items, and number of attributes accounrs over-all for less than 10% of the vadance. Indeed, in nearly all the pairs the lack of consistent findings is striking. In addition to supporting the introductory comments, such results also emphasize one of the major assumptions by Luria (1773) that seemingly similar or identical tasks may be processed by widely differing brain strategies. As a result, attempting to find a method for all similariries in general may be impossible as there is no such single method. It is to the authors' credit that they recognize some of the limitations of the model. From the perspective presented here, it can be suggested that this is not just the absence of some additional linear factors but is due to a fundamental error in the computer model. This is not to say that models of brain function are impossible bur rather chat such models need to reflect a more complex set of linear and non-linear and experiential factors which capitalize on higher mathematical theory and relationships. Only when such a modification is made in our approach will this research begin to approach the actual operation of the human mind

Cognition · Cognitive psychology · Cognitive science · Information processing · Linear model · Machine learning · Perspective (graphical · Point (geometry · Population · Similarity (geometry · Variance (accounting · Artificial Intelligence · Cognitive Science and Mapping · Computer Science · Neuroscience · Psychology

• A Cognitive Analysis of the Wais Similarities Task

  Open Access•Samuel C Fulkerson, Jerry L Cartwright•Psychological Reports•1984

  Add to list