Autonomy in the interactive music system VIVO Fabio Paolizzo 1,3,4, *, Colin G. Johnson 2 1 University of California, Irvine, Department of Cognitive Sciences, USA. 2 University of Kent, UK. 3 University of California, Irvine, Department of Dance, USA. 4 University of Rome Tor Vergata, Department of Electronic Engineering, USA. Abstract Interactive Music Systems (IMS) have introduced a new world of music-making modalities. But can we really say that they create music, as in true autonomous creation? Here we discuss Video Interactive VST Orchestra (VIVO) — an IMS that considers extra-musical information by adopting a simple salience-based model of user-system interaction when simulating intentionality in automatic music generation. Key features of the theoretical framework, a brief overview of pilot research, and a case study providing validation of the model are presented. This research demonstrates that a meaningful user–system interplay is established — what we define as reflexive multidominance. Keywords: interactive music systems, computational creativity, salience, reflexivity, multidominance. Tridimensional continuum: instrument, player, and score Rowe (1993) defines an IMS in terms of a continuum between instrument and artificial player. Here, we propose to extend this definition to include the score as an interdependent dimension of a three-dimensional space. This perspective mandates that an IMS (a) always affords an amount of control to the user via interactions with the instrument, (b) conveys system autonomy (i.e., suggests an autonomous player) to the user, and (c) constrains the invention scenario by its system architecture or score. We demonstrate the importance of this three-dimensional space by looking for commonalities among different systems for each specific dimension. Notably, by introducing the score as an explicit dimension of the continuum, we are able to investigate user interaction and system autonomy — a current limitation in IMSs. We hypothesize that, for an IMS, system autonomy exists within this space, and that it is possible to use extra-musical information (e.g., the quality of a gesture) in the generation of a sound in order for a user to identify the sound as being intentionally organized, which is a crucial step in interpreting music as meaningful. 1 Corresponding author Email addresses: fabio.paolizzo@uci.edu fabio.paolizzo@gmail.com (Fabio Paolizzo), C.G.Johnson@kent.ac.uk (Colin G. Johnson).