Lorenzo Magnani has pointed out the danger of considering scientific models not only as useful tools for explaining facts or discovering new entities, laws and theories, but also as fictions, surrogates, parables, etc. It is true that underlying many current positions is the inability to clearly distinguish between the epistemic and non-epistemic use of idealizations and/or fictions in the scientific enterprise. However, an account that primarily interprets idealization not at the level of theoretical modelling and thought experiment, but at the level of real experimentation and, more generally, human agency, can acknowledge the fictional character of empirical-scientific models and thought experiments while maintaining a firm distinction between empirical-scientific and fictional work.