On representation and similarity: The case of mouse models of cancer

According to the similarity account, scientists use models to represent their targets by utilizing similarities in certain respects and to certain degrees between a model and its target. According to the critics, however, representation is conceptually distinct from the notion of accurate representation, and rather than being the relation that grounds representation, similarity should be considered as setting a standard of accuracy. Based on the case study of research practices involved in using mouse models to study cancer, this paper argues that while the overarching skepticism regarding the similarity account may be justified, the role of similarity in specific contexts deserves attention. Indeed, it will be shown that similarity plays a significant role in determining whether a mouse model represents a particular aspect of cancer. Thus, authors dismissive of similarity grounding representation, while correct in the general picture, should take into consideration the role that similarity plays in deciding whether a model is or not a representation in concrete scientific practices.