Melody Recognition after Short Delays: Effects of Contour Complexity and Key-Distance

This study investigated melody recognition with a modification of the standard-comparison paradigm. Subjects listened to an original melody, were exposed to a silent retention interval, and then were presented with a target and a distractor. There were two types of trials. On target-same trials, listeners heard the target (the original melody) played in the same key exactly as it was previously heard and a distractor (a novel melody) played in a key either a major second or a perfect fourth from the original. On target-different trials, targets were heard in keys either a major second or a perfect fourth from the original, while distractors were played in the same key as the original. The contours of targets and distractors were also examined. Retention intervals varied from 0.5 s to 15 s. The results indicate that contour complexity and keydistance interact in the recognition of short melodies. An important goal in the field of music cognition is to understand how people perceive and recognize melodies. A melody is a succession of pitches in time. Melodies are encoded in an interval code and a contour code (Dowling, 1978; Edworthy, 1985). The interval code is related to the distance between any two adjacent tones in a melody along a logarithmic frequency scale. Intervals are the basis of musical scales and harmony. Contour, on the other hand, involves the pattern of up and down directions in a particular melody (Dowling, 1978; Edworthy, 1985; Massaro, Kallman, & Kelly, 1980). Together, interval and contour define a melody. Behavioral research has demonstrated that interval and contour are implicated in melody recognition (Cuddy & Lyons, 1981; Dowling & Bartlett, 1981; Dowling & Fujitani, 1971; Idson & Massaro, 1978; Massaro et al., 1980). Clearly, if we are to understand how melodies are perceived, we need to understand how interval and contour information are processed. Melodies can be transposed to different keys. Transposition occurs when a melody is shifted upward or downward in pitch while the intervals between pitches remain constant (Dowling & Bartlett, 1981; van Egmond & Povel, 1996). A transposed melody sounds similar to a nontransposed version except that it will be higher or lower in pitch. Transpositions can either be exact or inexact. An exact transposition retains the interval distance relationship between tones, whereas the interval structure in inexact transpositions is changed. When we hear familiar melodies, they are usually exact transpositions to arbitrary pitch levels (Dowling, Kwak, & Andrews, 1995). We recognize exact transpositions as the same melody, regardless of the key being played. This makes transpositions ideal for studying melody recognition. The distance of a transposition from an original melody can be measured in key-distance as well as pitchdistance. Key-distance is a measure of the extent that two keys share identical pitches. Key-distance is better understood if the circle of fifths is discussed. The circle of fifths describes a relationship between keys that shows that “close” (or near) keys have more notes in common with each other than “far” keys. If one takes a melody in C major for instance, the closest related key is G major. G major begins on the fifth note of the key of C and has only note that differs from C major, that of F#. The next key in the circle of fifths is D major, which begins on the fifth note of the G major scale. D major shares every note in common with G major except for C#. This progression continues as the fifth note of the current scale becomes the first note of the next scale through the entire cycle of fifths. Correspondence concerning this article should be addressed to Thomas W. Reiner, Department of Psychology, Troy University, 136 Catoma Street, Montgomery, AL, 36104. Email: [email protected].