Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)




Dr. J. R. Platt


When investigating the perceived similarity of two musical stimuli, systematic asymmetries emerge which depend on the nature of the two elements being compared, relative to the musical context in which they appear. At the level of individual tones, Krumhansl (1979) found that if the order of presentation was diatonic/nondiatonic (relative to a tonal context), similarity ratings were lower than if that order was reversed. She concluded that when a change results in an element becoming less stable in terms of its position on the tonal hierarchy, similarity perception will be lower than if a change increases its stability. At the level of melodies, Bartlett and Dowling (1988) obtained a similar result, but claimed the asymmetry was due to violations of scalar structure, having little or nothing to do with the tonal hierarchy. To test between these different accounts, a series of experiments was conducted in which pairs of diatonic melodies were presented for similarity ratings. Each melody consisted of a context sequence and a target sequence. The context sequences were designed to promote either a C-major or a D-minor tonal hierarchy. These respective keys share similar notes in their scales, but the tonal hierarchies are inverted with respect to one another. According to Krumhansl's account of the asymmetry, noticeable changes in one key context should not be noticeable in the other, due to the reversal in stability of the component tones of the alteration. According to Bartlett and Dowling, no differential sensitivity should be observed, since all changes were within the scale structure of both contexts. In the first experiment, repetition of the tonic was employed as the key-instantiating stimulus, the result being that strong asymmetrical perception arose, both on the measures of similarity ratings, as well as alteration detection ability. Subsequent experiments employed triadic contexts (suggested by Krumhansl and others to be strongly key-instantiating), and note-frequency controlling contexts (to rule out the possibility of note repetition playing a role in the similarity ratings). The results supported the hypothesis that asymmetric perception is a result of the dynamic tone quality differences between scale degrees in a tonal melody. Two subsequent control studies ruled out the possibility of target sequences themselves being responsible for the asymmetries, and confirmed that listeners perceived the melodies in the keys specified in the experiment. A model based loosely on Tversky's explanation of asymmetric perception was put forth to explain these data, as well as those of Bartlett and Dowling.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Included in

Psychology Commons