Thesis (M.Ed.)--Chulalongkorn University, 1980

Two inter-supporting experiments were specially designed for investigation on natural mechanism of speech perception, focused at the economy of encoding process for noun and modifiers. In one experiment based upon displacement hypothesis, verbal presentation of varied-modifier positions of twenty four patterns, each followed by a string of random digits exceeding memory span, were assigned to each one of twenty four subjects. Subjects were required to perform immediate serial recall verbally after presentation of every pattern. In another experiment, based upon differential processing time hypothesis, reaction time of verbal-visual matching of the varied-modifier-position items and pictures was in consideration of differential processing time of the encoding process. A synchronized system of electronic apparatus was constructed and each on of another group of twenty four subjects performed the verbal-visual matching for twenty four patterns corresponding to the first experiment. Data from both experiments were analyzed by means of ANOVA, conditional error probability and position reversion analysis. The results illustrates differential effect of modifier positions upon perception and short-term memory. Furthermore, positional recall probability and position reversion analysis suggest a reorganizing mechanism which can be considered as a syntactic reorganization or a perceptual unit reorganization in which the acoustic codes are processed with a limited capacity of 7±2 units (chunks). Inevitably, the reorganization of the codes is operated by the derived strategies and techniques from long-term memory. Supporting by the results of conditional error probability, it is apparent that the error results from the failure of instantaneous chunking; therefore, the set of codes are impelled into the reorganization loop. The more times the looping occurs, the more probable the interference occurs. Thus speed and efficiency of encoding process decrease as a linear function of the number of modifiers prior to noun in the right-hand-branching language. Further experiment on the left-hand-branching language for supportive data is recommended. Finally, relative to the contemporary theories of speech perception and human memory, these findings conceivably bring about a complete model of speech perception. The proposed model is discussed delicately in many aspects and further experimental verification by neuropsycholinguistic approach and computer simulation is recommended.