The cohort model of melody identification: Evaluating primacy and similarity

Abstract

Two experiments investigated two hypotheses derived from the cohort model of melody identification. The primacy hypothesis holds that the initial notes of a melody are differentially important for identification because they are used to activate the cohort. The similarity hypothesis holds that the likelihood of identifying a stimulus as a particular song will depend on the musical similarity of the stimulus to that song. Consistent with the primacy hypothesis, identification performance decreased significantly when the initial notes of the melody were not heard. However, musical similarity was generally a poor predictor of identification performance and confusion errors were rare and showed little evidence of musical similarity. These data are used to argue that melody identification does not involve an initial activation of candidate melodies based on low-level musical properties that are gradually winnowed down on a note-by-note basis. Instead, melody identification seems to be more ‘all-or-none’ than the cohort model suggests.