On the Music of Emergent Behavior: What Can Evolutionary Computation Bring to the Musician?

Abstract

In this article, the author focuses on issues concerning musical composition practices in which emergent behavior is used to generate musical material, musical form or both. The author gives special attention to the potential of cellular automata and adaptive imitation games for music-making. The article begins by presenting two case-study systems, followed by an assessment of their role in the composition of a number of pieces. It then continues with a discussion in which the author suggests that adaptive imitation games may hold the key to fostering more effective links between evolutionary computation paradigms and creative musical processes.

In this article, I forgo the discussion of whether or not computers can compose music, seeing it as no longer relevant: computers can compose, if programmed appropriately. Perhaps one of the greatest achievements of artificial intelligence (AI) to date lies in the construction of machines that can compose music of high quality, such as the Experiments in Musical Intelligence (EMI) system [1]. However, these AI systems are either hard-wired to compose in a certain style or able to learn how to imitate a style by looking at patterns in a group of training examples; they do not create new musical styles, so to speak. Conversely, issues of whether computers can create new kinds of music are much more difficult to study because, in order to test this idea, the computer should neither be embedded with particular models at the outset nor learn from carefully selected examples. Furthermore, it is difficult to judge what the computer creates in such circumstances, because the results normally sound very strange to us; they tend to lack the cultural references that we normally rely on when appreciating music.

One plausible approach to these problems is to program the computer with abstract models that embody our understanding of the dynamics of certain compositional processes. Since the invention of the computer, many composers have tried out mathematical models thought to embody musical composition processes, such as combinatorial systems [2], stochastic models [3] and fractals [4]. Some of these trials have produced interesting music and much has been learned about using mathematical formalisms and computer models in composition. The potential of evolutionary computation is, therefore, a natural progression for computer music research. In this article, I introduce a discussion on the potential of cellular automata and adaptive games for the composition of truly "new" music (albeit music that has the potential to make sense to our ears). The article begins with a brief introduction to CAMUS (Cellular Automata MUSic) and Chaosynth, two cellular automata-based music systems of my own design, followed by an assessment of their role in the composition of a number of musical pieces.

Case Studies

CAMUS uses two simultaneous cellular automata (CA) to generate musical forms: the Game of Life (Fig. 1) and Demon Cyclic Space. Due to limitations of space, I will briefly introduce only the role of the Game of Life in the generative process [5].

Propagating Cellular Automata Musical Forms

The Game of Life is a two-dimensional CA that attempts to model a colony of simple virtual organisms. In theory, the automaton is defined on an infinite square lattice. For practical purposes, however, it is normally defined as consisting of a finite m × n array of cells, each of which can be in one of two possible states: alive, represented by the number one, or dead, represented by the number zero. On the computer screen, living cells are colored as black and dead cells are colored as white (see Fig. 1). The state of the cells as time progresses is determined by the state of the eight nearest neighboring cells. There are essentially four rules that determine the fate of the cells at the next tick of the clock:

1. 1. Birth: A cell that is dead at time t becomes alive at time t + 1 if exactly three of its neighbors are alive at time t

2. 2. Death by overcrowding: A cell that is alive at time t will die at time t + 1 if four...