The aim of the present study was to investigate how context effects, p
reviously observed in a perceptual task (Bigand, 1993, in press), infl
uence melody recognition. Three factors were manipulated: key context,
rhythm context, and musical expertise. The key context effect was ach
ieved by making a few small changes in the pitches of the melodies, an
d the rhythmic context effect was achieved either by shifting the rhyt
hmic structure or altering the duration of the tones. Participants wer
e asked to evaluate the percentage of tones that had been changed from
a standard to a comparison melody. The data support the dynamic theor
y of attention developed by Jones, (1976), Jones and Boltz (1989), and
Jones and Yee (1993). The perception and memorization of melodies are
dynamic, context-specific activities: irrespective of musical experti
se, varying the strength of tonal and rhythmic accents in a melody aff
ects the way listeners' attention is guided through time, which in tur
n alters the overall dynamic shape of the melody and renders its recog
nition difficult.