Interactive reflexive musical systems for music education

Our project collects and develops experiments about music education and the Interactive Reflexive Musical Systems. The IRMS are systems in which the user, whatever his competences’ level, is confronted with a kind of developing mirror of him/herself. The Continuator is a particular kind of IRMS able to learn constantly by the input is given and to produce music in the same style as a human playing the keyboard. The phrase generated by the system is similar but different from those played by the users (Pachet 2006). The system was elaborated for professional musicians. We decided to experiment it with children. In particular, we chose to study young children, 3/5 years old, because at this age the problem of the interaction between child and machine takes on a fundamental role in the learning process. We carried out an experimental psychological protocol and several classroom setting experiences. The results suggest this system is able to develop interesting child/machine interaction and creative music processes in young children, creating a state of well-being very similar to this described in the Theory of Flow by Csikzsentmihalyi, enhancing attention span, self-regulation and self-practising. In the group setting it is a versatile device to support music teaching and collaborative learning. The innovative function of the Continuator is the creation of a dialogue with the child, while he is playing the keyboard. Between the system and the child a circular interaction is setting up, in which the child’s musical style influences the system, which answers repeating and changing the child’s musical fragment, in a continuous improvisation process. This dialogue is not only founded on the turn-taking rules, but also on the mechanism of repetition and variation similar to one observed in the infant/mother interaction. On the base of these results, other experiences started in Italy and other countries (UK, USA) with the aim to create a spiral collaboration between system designers and experts in pedagogy and psychology. The next contribution will be to propose novel systems which will benefit from the analysis of the new experiments. We are strongly interested in the application of the IRMS concept to the content-based music indexing, as developed in the MusicBrowser project at SONY-CSL, to facilitate the access of collections of existing music repertoire for children.