SummaryThe main argument for scientific realism is that our present theories in science are so successful empirically that they can't have got that way by chance ‐ instead they must somehow have latched onto the blueprint of the universe. The main argument against scientific realism is that there have been enormously successful theories which were once accepted but are now regarded as false. The central question addressed in this paper is whether there is some reasonable way to have the best of both worlds: to give the argument from scientific revolutions its full weight and yet still adopt some sort of realist attitude towards presently accepted theories in physics and elsewhere. I argue that there is such a way ‐ throughstructuralrealism, a position adopted by Poincare, and here elaborated and defen