Abstract

Physics appears to be in a unique position to withstand antirealist attacks, especially ‘pessimistic induction’ arguments. Such resilience provides an incentive for embracing a ‘structural’ variant of scientific realism. Nevertheless, an examination of the physics-mathematics relationship suggests that whatever determines the success of modelling endeavours lends scant support to structural realism. A closer look at conceptual prerequisites of joint ab initio derivations of (i) Galilean and ‘special’ relativity theories and (ii) classical and ‘quantal’ probabilistic frameworks also fosters scepticism about a major claim of structural realists: that mathematical features of theoretical frameworks – those, in particular, which determine their predictive efficiency – are ultimately rooted in ‘structural’ aspects of the natural world. Further concerns with the central tenets of structural realism are finally briefly addressed, and the question of possible non-instrumentalist alternatives is raised.