THE PERCEPTION OF TONAL RELATIONSHIPS IN MUSIC (TONALITY, RHYTHM, ATTENTION)

The effects of rhythmic context, starting tone, and inclusion of a tritone interval on judgments of the tonality of short melodies were examined in three experiments. Musically sophisticated listeners indicated the tonality of melodies by reporting the scale step of the last tone in the melodies. None of the variables produced effects in the first experiment in which rhythm was treated as a within-subject variable and only one basic tone sequence was presented. However, all of the variables produced effects in the other experiments wherein rhythm was treated as a between-subjects variables and a greater variety of melodies and keys were presented. The pattern of accenting and metrical organization associated with a given rhythm was a major factor in determining the tonal structure that was detected by the listener. Effects of rhythm were largely due to the rhythmic accent placed on a tone at the beginning of a melody: Subjects tended to perceive the first accented tone as the tonic. In addition, rhythm appeared to influence tonality judgments by directing attention to pitches that outlined common chord progressions in different keys. These effects of rhythm indicate that the perception of tonal relationships in music can be understood only if we take into account the relationship between the pitch structure and the rhythmic structure of a musical stimulus. However, rhythm effects were greatly diminished when a tritone was included in the melody. Subjects appeared to be quite sensitive to the fact that the tritone (plus any other tone) is uniquely identified with a particular diatonic set. They almost always judged the melodies to be in the key that was indicated by the tritone. Nevertheless, the effectiveness of the tritone as a tonal indicator was influenced by the rhythmic context in which it appeared. This pervasive influence of rhythm, in addition to having implications for the perception of tonality, suggests that theories of attention need to account for the role of temporal structure in directing attention to particular stimulus relationships within patterned events.