Influence of surface features on the perception of nonadjacent musical phrases

Although temporally nonadjacent key relationships (e.g., Key X →Key Y→ Key X) are ubiquitous within tonal music, the full extent to which they are perceived is uncertain. Previous research suggests that memory for an initial key remains active up to 20 s after modulation; however, homophonic textures were used in these studies, leaving open the possibility that surface features such as figuration may contribute to nonadjacency effects. In two experiments, we investigated this issue by measuring goodness of completion ratings for stimuli in which musical surface features were manipulated. Two types of surface feature were tested: figuration and activity (total number of notes per stimulus). Stimuli were composed of three parts: (1) nonadjacent section (in either the same or a different key to the probe); (2) intervening section (in a different key to the probe); and (3) probe (a cadence in either the same or different key as the nonadjacent section). In Experiment 1, we tested whether the presence of surface features resulted in higher goodness of completion ratings for the probe; in Experiment 2, we manipulated nonadjacent key relationships to ascertain the effect of surface features on global perception of key. Results showed that figuration and activity contributed to goodness of completion ratings, particularly in stimuli where these features matched each other in the nonadjacent sections. Moreover, the presence of surface features strengthened the perceived relationships between the keys of nonadjacent sections, thereby appearing to contribute to the global perception of phrase. In sum, although from an analytical perspective surface features are often considered to be less important hierarchically, our results indicate that they contribute significantly to the perception of nonadjacent key relationships.