More about the (weak) difference between musicians' and nonmusicians' abilities to process harmonic structures
Bigand, E.*, Poulain, B.*, D'Adamo, D.*, Madurell, F.** & Tillmann, B.***
* LEAD-CNRS, Université de Bourgogne, France
** Music Department, Université Paris IV Sorbonne
*** Dartmouth College
Musicians and non musicians often behave similarly when they are required to perform musical tasks that are not more familiar for the former than for the later. The purpose of the present study was to further investigate this issue by using an on-line experimental paradigm designed to assess the influence of cognitive and sensory components on the development of harmonic expectancy.
Experiment 1 involved a harmonic priming task in short contexts. The aim was to assess the contribution of musical expertise on the differentiation between regular and less regular resolutions of a diminished chord. In one condition, a prime chord (say a B diminished chord) was followed by a target that was either one of the four possible resolutions of the diminished chord (C major chord for example) or a less legal resolution (C# for example). Interestingly, less legal targets share more component tones with the prime than do legal targets. According to sensory priming, illegal targets should be processed more easily than legal ones due to the shared component tones. According to Western musical rules, the processing of the target chord should be more facilitated (more accurate and faster) for legal targets than for illegal targets. The critical point of the study was to investigate the contribution of sensory and cognitive components of harmonic priming as a function of the extent of musical expertise. Sensory priming was expected to predominate over cognitive priming in non musicians, and a reverse tendency was expected in musicians.
Method. Participants performed a harmonic priming task by providing a simple perceptual judgment on the target chord. Following Bharucha and Stoeckig (1987), the required perceptual judgment concerned the sensory consonance of the target. For the purpose of the experimental task, half of the target chords were rendered dissonant by adding either an augmented octave or an augmented fifth to the perfect major triad. The velocity of this added tone was adjusted in order to render the dissonance moderately hard to perceive. Participants had to judge as quickly and as accurately as possible whether the second chord of the pair was consonant or dissonant. This perceptual judgment do not require participants to pay attention to the harmonic relationship between prime and target. However, we expected that this judgment will be more or less easy depending on this harmonic relationship.
Results. Participants with high level of musical training and participants with no musical training processed the target chord more easily in the legal resolution condition than in the illegal condition. This outcome suggests that for both musicians and non musicians cognitive components predominate over sensory components in a harmonic priming.
The goal of Experiment 2 was to extend this conclusion to longer harmonic contexts. Harmonic priming has been shown to occur in chord sequences of different lengths for both musicians and non musicians. In Bigand & Pineau (1997), the target chord was easier to process when it acted as a tonic rather than as a subdominant chord in the preceding context. Bigand and Pineau's experiments were not designed to definitely contrast the respective contribution of sensory and cognitive priming. The chord sequences used in our present study attempted to address this issue. In two conditions (no sensory priming conditions), the target chord never occurred in the preceding context. As a consequence, the context never contained a tonic or a subdominant chord. Nevertheless, we expected a stronger priming effect on the (implied) tonic chord than on the (implied) subdominant chord. In two further conditions (sensory priming conditions), the subdominant chord occurred one or two times in the context. The prior occurrence of the subdominant chord should increase the influence of sensory priming. If long chord priming primary depends on a sensory component, the processing of the subdominant chord should be easier than the processing of the tonic target chord. If long chord priming primary depends on a cognitive component, the processing of the subdominant chord should remain more difficult than the processing of the tonic target. Once again, the critical point of the study was to assess the influence of these sensory and cognitive components as a function of the extent of musical expertise.
Method and results. The method was identical to that of Experiment 1. The results demonstrated a strong harmonic priming effect for the tonic chord. This priming effect was strictly unchanged when comparing the sensory priming conditions to the no sensory priming conditions. Once again, non musicians showed highly similar patterns of results.
Following the same rationale, we investigated in Experiment 3 the influence of horizontal motion on the processing of the target chord. There was main effect of horizontal motion (with target chords being more difficult to process in the bad voice leading condition), but no influence of horizontal motion on harmonic priming. Irrespectively of the voice leading, tonic target chords were easier to process than subdominant target chords. This finding was observed for both musicians and nonmusicians.
Conclusions. These experiments provided evidence that for both musicians and non musicians the processing of subtle changes in harmonic structure involves a sophisticated cognitive component that does not depend on the extent of musical expertise.
Bharucha, J. J. & Stoeckig, K. (1987). Priming of chords: Spreading activation or overlapping frequency spectra? Perception and Psychophysics, 41, 519-24.
Bigand, E., & Pineau, M. (1997). Global context effects on musical expectancy. Perception and Psychophysics, 59, 1098-1107.
This research was supported by the International Foundation for Music Research.
Back to index