Priming of chords: Spreading activation or overlapping frequency spectra?

Abstract

A chord generates expectancies for related chords to follow. Expectancies can be studied by measuring the time to discriminate between a target chord and a mistuned foil as a function of the target's relatedness to a preceding prime chord. This priming paradigm has been employed to demonstrate that related targets are processed more quickly and are perceived to be more consonant than are unrelated targets (Bharucha & Stoeckig, 1986). The priming experiments in the present paper were designed to determine whether expectancies are generated at a cognitive level, by activation spreading through a network that represents harmonic relationships, or solely at a sensory level, by the activation of frequency-specific units. In Experiment 1, prime-target pairs shared no component tones, but related pairs had overlapping frequency spectra. In Experiment 2, all overlapping frequency components were eliminated. Priming was equally strong in both experiments. We conclude that frequency-specific repetition priming cannot account for expectancies in harmony, suggesting that activation spreads at a cognitive level of representation. © 1987 Psychonomic Society, Inc.