I asked the question at the beginning of this section on Sundanese gamelan music: how is it possible for a group of musicians to play highly complex music, in a cohesive manner, without the use of notation and without having to memorise impossibly large amounts of music? My answer came in a number of stages.
Rather than reading, or memorising vast amounts of music, the musicians memorise the simple frameworks of pieces (the Javanese term for this, balungan – literally, skeleton – is sometimes used by Sundanese musicians). Individual parts are all related to these frameworks.
Parts are worked out by the application of simple principles or rules, which allow for several equally valid realisations. Having learned their instrument's part for a few pieces, a musician can apply the same principles to other pieces too, working out the part as he goes along – in performance.
A strong sense of ensemble ensures a cohesive performance. Everyone knows, at least in general terms, how the other instrumental parts go: musicians listen to each other and respond to each other's changes. In particular,
(a) in the case of some instruments (such as the sarons), separate parts interlock to form a coherent whole;
(b) in order to expand and contract the piece together, everyone must listen and respond to aural signals given by the drummer.
To bring this back to the discussion at the beginning of the course, everything I've said about Sundanese gamelan relates quite clearly to Nettl's description of music which is ‘spontaneous but model-bound, rapidly created, and simply conceived’ (see Section 1.2). Somehow neither the term ‘composition’ nor the term ‘improvisation’, as generally understood in the West, quite captures the way this music is created. It should be clear by now, as I have suggested all along, that this dichotomy is a very limited tool for describing musical creation.