Short-term auditory memory interference: The Deutsch demonstration revisited

The interference effect of nonspeech and speech in short-term auditory memory was investigated with an experiment paradigm proposed by Deutsch [Science, 168, 1604-1605 (1970)], in which test tones, separated by a 5-s retention interval, were interpolated with six other sounds. In Experiment 1, the test tones were pure tones. The interpolated sounds were pure tones and naturally spoken digits by a female and a male. Nine participants were tested for (1) pure-test-tone pitch recognition, (2) serial recall of the interpolated spoken digits, and (3) both tasks (1) and (2). Pitch recognition errors were significantly increased in task (3) compared to task (1), and the digit recall errors were also significantly increased in task (3). In Experiment 2, the test tones were eightcomponent harmonic complex tones. The interpolated sounds were eight-component harmonic complex tones, and naturally spoken digits by a male. Twelve participants were tested for the corresponding task conditions as in Experiment 1. Significant increases in the errors of pitch recognition and of digit recall were observed when both tasks were required. These results suggest that speech can interfere with tone pitch in short-term auditory memory, and that pitch salience plays a crucial role in the interference.