Abstract
In this paper I juxtapose two well-known thought-experiments concerning duplicate art works, and point out that they appear to have directly conflicting results. I then make a proposal as to how to reconcile the two cases. The two cases are Borges' story of Pierre Menard, in which a text coinciding exactly with Cervantes' Don Quixote is nonetheless a distinct work from it, and Nelson Goodman's claim that a musical work cannot be forged, because anything complying with a work's notation is that work (or a performance of it), not anything distinct. Both claims seem correct, but how can they be reconciled? I suggest two factors that must be present for two indistinguishable works to be numerically distinct: divergent causal histories and distinct 'privileged interpretational instructions'. Notational equivalence of W and W1 plus the instruction to interpret W as W1 leads to the impossiblity of forgery. But in the absence of such instruction, W and W1 can constitute distinct works.