This paper criticises the Waltonian fiction view for providing a misleading role of imagination in scientific modelling, and for failing to provide an adequate account of the epistemology of modelling. Imagination cannot be simultaneously constrained by the model descriptions and relevant for modelling epistemology. Given that the relevant inferences must be made in terms of publicly available model descriptions, and the laws and general principles must be included in the model descriptions, there can be no relevant role for the so-called indirect principles of generation in modelling epistemology.

Large-scale numerical simulations are increasingly used for scientific investigation; however, given that they are often needed precisely because ordinary experimental and observational methods cannot be used, their epistemic justification is often in question. Drawing on the adequacy-for-purpose framework, I characterize the problem of model assessment under conditions of scarce empirical evidence. I argue that, while a single model may not suffice under these conditions, a suitable collection of models may be used in concert to advance a community's scientific understanding of a target phenomena and provide a foundation for the progressive development of more adequate models.

The cultural red king effect occurs when discriminatory bargaining practices emerge because of a disparity in learning speed between members of a minority and a majority. This effect has been shown to occur in some Nash Demand Game models and has been proposed as a tool for shedding light on the origins of sexist and racist discrimination in academic collaborations. This paper argues that none of the three main strategies used in the literature to support the epistemic value of these models—structural similarity, empirical confirmation, and how-possibly explanations—provides strong support for this modeling practice in its present form.

We present a computational model of reflective equilibrium with precedent. Each agent considers a rule by which to accept or reject cases. Cases are represented as labeled binary strings: intuitive accept, intuitive reject, or no intuition. Rules are represented as a pair: a binary string and a tolerance threshold determining if a case is a close enough match to accept. Rule-updates are driven by intuitions about cases and precedents set by other agents. We compare four networks: empty, ring, 4-regular, and complete. Results suggest that increasing connectivity encourages, but doesn't guarantee, interpersonal convergence on a single reflective equilibrium.