On the Redintegrative Perception of Words

Abstract

According to Solomon and Postman, When a stimulus-pattern is presented at short durations or at low illumination-intensities, only fragments of the total wordstimulus are 'effective.' Such a stimulus-fragment may be considered to represent a point on the generalization-dimension of stimulus-patterns capable of eliciting the correct verbal response.' Certain questions regarding this statement immediately present themselves: (1) Where in the word are the stimulus-fragments that are the effective cues? (2) What conditions attend the redintegrative leap from the fragment to the correct whole-response? A partial answer to the first question is given by Erdmann and Dodge, Wagner, and others who found in nonsense arrangements of letters that the first and last letters were often correctly recognized while the middle letters were not.2 This finding needs to be checked for meaningful words. Also only with meaningful words can the second question be answered. Words of length greater than the individual's span of apprehension for letters were chosen as the exposure-material and were so chosen that the responses ranged from correct to fragmentary reports. Ever since Cattell's experiments, the crux of the problem seems to be why familiarity and meaning make it possible for a S to surpass his span for unconnected letters.8 One may experimentally induce such superiority by 'building in' familiarity, as in Solomon and Postman's experiment, or by using extraneous organization, as in Vanderplas' study.4 A direct, immediately available measure of familiarity can, however, be found in the individual's ability to define a word. Are words, as equated in Thorndike's frequency of usage, perceived more readily in short exposures when S can define them than when he can not? Does the ability to define have any relation to the ability to recognize the stimulus-word among alternatives presented after an unsuccessful recognition? The study of the function of various aids to recognition may provide a solution to Cattell's problem.