The process of learning music is an integral part of a music culture, but this process can be transformed across cultural borders by combining cultural representations using visual, gestural, auditory, linguistic and spatial modes. Intercultural arts praxis is discussed in relation to traditional indigenous song and dance in the Torres Straits of Australia and its relevance to contemporary performance occurring in classroom music improvisation in an inner city Australian school. Decision-making during performance reflects goals, interests and values of performers in praxis. This paper uses multimodal visual analysis to examine aspects of the pedagogic process and the making of meaning in context. In video transcripts, presentation of many modes simultaneously occurs as symbolic representations are added to a music score. Such visual representations allow for new, heightened possibilities of meaning. As exemplars of voices of artists, audiences, theorists and children, the visual transcripts enliven research in the intercultural arts. The consideration of the relationship between in-school and out-of-school practice in the development of knowledge is made as the researcher introduces 'the space of music dialogue.' Linked to social critical theory, this conceptual framework facilitates the viewing of multimodal interactions, and relations between local and global cultural practices in the art