relations (see below)). Thus despite its fundamental role in melodic perception, parallelism
and its perception is only beginning to be investigated in a quantitative framework.^20
Intervallic implications
Listeners are sensitive to the patterning of successive pitch intervals in a melody. One
source of evidence for this is that if a melody is stopped midstream, listeners typically have
well-defined expectations of how it will continue.^21 Interval-based expectancies are the
focus of Narmour’s ‘implication-realization’(IR) model.5,6A flavour for the model’s prin-
ciples can be gleaned from Schellenberg,^22 who investigated melodic expectancy in British
and Chinese melodies using American and Chinese listeners. Two principles emerged as
important factors cross-culturally:‘pitch proximity’(an overall expectancy for small
intervals) and ‘pitch reversal’(an expectation that after a large interval, the pitch direction
will change). The source of these bottom-up principles may be auditory Gestalt principles^23
or experience with the intonation contours of speech.
Tension vs resolution
A remarkable aspect of melody perception is an unfolding sense of tension vs resolution
over the course of the entire melody. That is, at each point a listener has some sense of
whether the melody must continue or could naturally come to a halt. Bigand^24 has empir-
ically measured this aspect of melody by having participants listen to increasingly long
fragments of a melody and judge the stability of each fragment’s end.eHis results confirm
that perceived tension varies widely during the course of a single melody. This is easily
illustrated with K0016: consider two fragments which stop at the ends of phrase 3 and 4,
respectively (sound examples 3 and 4). These differ dramatically in the degree of tension
experienced at the final tone, even though both fragments stop at the ends of phrases.
Two factors that contribute to the degree of perceived tension are ‘tonal weight’and tone
duration.^24 The latter is simple (longer toneless tension), but the former requires a brief
explanation. In ‘tonal’music of the Western European tradition different pitches take on
distinct structural roles in the fabric of the music. Consider Figure 21.4, in which each note
of K0016 has been marked with a number representing each tone’s position (or ‘scale
degree’) in the scale from which the melody is built (e.g. tone 1 is do, 2 is re, 3 is mi, etc.).
Music perception research shows that once a melody is perceived as being in a particular
key, certain scale degrees act as perceptual reference points,^25 especially scale degree 1 (also
known as the ‘tonic’). K0016, for example, is in the key of C, and the central role of the
C note (scale degree 1) is marked in several ways. It is the final tone of phrases 1, 2, 4, and 5,
and thus serves as a resting point within the melody and at the melody’s end. Furthermore,
it always occurs on a beat and is always long in duration. Thus scale degree 1 acts as a per-
ceptual ‘centre of gravity’for the entire melody, a role which can be contrasted with that of
another tone in the melody (D), marked by the numeral 2 in Figure 21.4. Tone 2 is far more
329eBigand (^24) (p. 811) defined high stability at the end of a fragment as ‘the feeling that the melody could natu-
rally stop at this point...[while] low stability [is] the feeling that there must be a continuation of the melody’.
Listeners rated stability on a scale of 1–7. My use of‘tension’is equivalent to Bigand’s ‘low stability’. Bigand^24
reported stability profiles, which can easily be converted to tension profiles by taking their inverse.