Any attempt to divine the logic of generative music must begin with the idea of propagation—how a minute element can proceed unchecked, ramifying unpredictably across a structure. In the context of Max/MSP, one realises this in practice through probability weighting within the random object, or by modulating parameters with a [drunk] walk that lurches, sometimes erratic, sometimes inevitably onward. The composer here becomes less an author than an overseer: a watcher of small initial sparks that, when left to their own algorithms, achieve domains of textural complexity lying beyond intention.

This process is not one of linear growth, but of thresholds: a carefully balanced generative patch can, under miscalculation or excessive feedback, tip into cacophony, uncontrolled proliferation, a sonic conflagration. Consider how Brian Eno’s early generative systems—say, in ‘Music for Airports’—eschew wild recursion, instead delineating soft boundaries: delays with defined feedback percentages, probabilistic note choices clamped to modal harmonics. To abdicate all oversight would be to risk aesthetic collapse, the hubris of unbounded emergence.

There is a lesson, then, in the interplay of freedom and constraint within generative systems. The art, such as it is, consists in tuning the conditions under which the unexpected may flower. Musicians nudge their code or modular apparatuses to the brink—reinserting chance, but circumscribing risk with gates, limiters, and probability distributions. Rather than setting a conflagration, they master the slow burning: a music that glows, mutates, and persists at the lovely edges of its chosen design.