Paths in the Forest of Signs. Concluding remarks — graphic variation, commutation, and the two inventories

Paths in the Forest of Signs draws together the conference's contributions on graphic variation across the world's writing systems—deciphered, partly deciphered, and undeciphered. It confronts the founding question: does the identification of allographs presuppose decipherment, or can a disciplined analysis of graphic variation precondition it? The apparent contradiction dissolves once one principle, shared in practice by every speaker, is made explicit: the identity of signs is never read upward from visual resemblance but established downward, by commutation. The pertinent question is therefore not temporal but structural—to which inventory does a given visible difference belong? We argue that every script keeps two complementary inventories: a graphemic one (graphs, organised by allography, suballography and allographemy, structured by the relation to language, and the object of decipherment) and a grammemic one (grams, bearing symbolic, indexical and iconic values, far more open and fluctuating). A single glyph instantiates both at once—which is why no single hierarchy can order them, and why the recurrent quarrel over how many signs a script contains is a symptom of conflating them. Allograph identification thus neither precedes nor follows decipherment: it is one commutational operation, run prospectively where we cannot yet read, retrospectively where we can.