The Challenge of Miscomputation

Explaining miscomputation is often a central desideratum of accounts of physical computation. Our claim is that explaining miscomputation is a non-trivial task. Specifically, miscomputation poses a challenge for norm-based accounts of computation. If the system's norm is to compute a specific function, in deviating from this function, the system does not comply with the norm, and hence, does not compute at all. If the norm is to compute some (non-specific) function, in computing any function the system complies with the norm; and hence, it computes but does not miscompute. On either horn, norm-based accounts face the problem of explaining miscomputation. We describe and motivate the dilemma and examine some possible strategies for avoiding it.