Composers must produce new music that neither repeats existing music, nor may not be construed by connoisseurs as a close variation of existing music, according to cultural standards of originality. Despite the enormous burden thus placed on memory and invention, non-derivative composers (here, J.S. Bach) never seem to inadvertently replicate existing music in part or whole: appropriation, probably conscious, implies improvement. The problem is how the composer's memory may be organized and managed in order to to accomplish this.
Memory operation can be witnessed in the creative problem of attaining distinctiveness in sets of works. Distinctiveness implies differential treatment of especially salient shapes that occur in all music. We argue that the most salient shapes are necessarily the simplest, and that an emergent structure of simplicity, in the sense of Elizabeth Bates, is available through intuition rather than cultural transmission. This is modeled as a recursive shape grammar whose base elements are the enumerable set of simplest things that can be done with musical material. The simplest class of shapes so obtained are easily demonstrated to exist in all musical cultures. Nevertheless, no historical rubric exists for this category: the creative management of simplicity may well be intuitive. This strengthens the implication that there is a parallel between compositional memory and the organization of patterns promoting difference.
Applied to the solo violin and cello music of J.S. Bach, our model reveals robust _avoidance_ of similarly-patterned simple structures in a combinatorial matrix governing bodies of works that are heard as unique and distinct. This result offers an argument that compositional memory is procedural rather than semantic: rather than remembering music as a succession of individual notes, music is remembered as structures which implicitly offer dynamic potential for difference and variation, much as chess players view configurations of the chessboard.