Aggregate context effects in music processing

Control of stimulus confounds is an ever-present, and ever-important, aspect of experimental design. Typically, researchers concern themselves with such control on a local level, ensuring that individual stimuli contain only the properties they intend for them to represent. Significantly less attention, however, is paid to stimulus properties in the aggregate, aspects that, although not present in individual stimuli, can nevertheless become emergent properties of the stimulus set when viewed in total. This paper describes two examples of such effects. The first (Case Study 1) focuses on emergent properties of pairs of to-be-performed tones on a piano keyboard, and the second (Case Study 2) focuses on emergent properties of short, atonal melodies in a perception/memory task. In both cases these sets of stimuli induced identifiable tonal influences despite being explicitly created to be devoid of musical tonality. These results highlight the importance of monitoring aggregate stimulus properties in one's research, and are discussed with reference to their implications for interpreting psychological findings quite generally.