On the Development of an Associative Response Time (RT) Measure for Semantic Satiation

Abstract

Loss in intensity of meaning of a word following its massed evocation has been measured through ratings on the semantic differential (Lambert & Jakobovits, 1960). To study such behavior in children and mental retardates who cannot make meaningful ratings on these differential scales, an alternative and relatively direct measure like associative RT is needed. According to existing notions (Lambert equates satiation and cognitive I,:) and experiments (Paul, 1962), RT for satiated words should be longer than for nonsatiated words. The experimental procedure included learning of 10 highly associated wordpairs, 5 of which were later given a standard satiation treatment of individual repetition of each word for 40 sec. at the rate of 2 or 3 per sec. The remaining 5 were treated as controls. Half of the 40 female undergraduate Ss were given the stimulus item for massed repetition, the other half received the response item for repetition. In order to obtain a measure of associative RT differences between pre- and post-repetition latencies, the list of 10 pairs was first presented 5 times for anticipatory recall on recorded tape and latency of response was recorded by a combined device of voice key and electrical timer. The entire list was again presented, using the same device, for 2 trials following massed repetition of 5 of the words. From the mean RTs for the last 2 pre-repetition trials, the mean for the 2 post-repetition trials was subtracted, separately for the satiated and control words. Subsequently, the control score was subtracted from the score for satiated items. This final RT score indicated the extent of shift in RT due to satiation. Results of a Lindquist Type I analysis of variance did not show a significant main effect for satiation treatment. Thus, the effect of satiation on RT, if any, would be very small. However, the final RT score and the semantic differential rating scores of each S correlated -.376 (rho, N = 20, P = .05) only in the response-satiated group. This implied, contrary to expectations, a facilitation in response speed with increased satiation. However, the relation is so small that it can at best raise doubts as to the efficiency of associative RT as a substitute for semantic differential ratings. The latter yielded a moderately strong satiation effect. Mean polarity difference score was -5.75 over 5 words and 9 scales.