Figure 8 Examples of the interaction between the tonal weight and the duration (short/long) observed on ending unit 7 (a), and between the tonal weight and the metrical position (weak/strong beat) observed on ending unit 12 (b) with musician subjects.

To summarise, the musicians’ results are consistent with those of the first experiment and they give more information about the interaction between implicit harmony×rhythm.

It is not the same when we look at the non-musicians’ results. These new profiles differ radically from those of the first experiment: they are more

contrasted, and more opposite in each experimental situation generated. The multivariate analysis of variance points to an effect of the implicit harmony significant at p<.001, which is the most important difference between the two experiments. Here, each musical progression generates contrasted tensions and relaxations schemas. Differences between them are qualitatively the same as those observed with musicians (see exp. 1). An interesting phenomenon is observed again for the ending unit 12. The average degree of musical stability is 3.33 for the E in H1, and 4.39 for the E in H2. This difference, significant at p<0.03, proves that even non-musician subjects are able to abstract very subtle differences in musical function.