MARCUS RENATO ALVES ARAÚJO

In the context of the debate between scientific realists and antirealists, it is reasonable to assert that the thesis of the underdetermination of scientific theories by evidence, alongside the pessimistic meta-induction and the criticism of inference to the best explanation, constitutes one of the main arguments against the explanatory defense of scientific realism. While various versions of this thesis exist in the literature, the issue of underdetermination generally emerges as a challenge of a broad nature to the rationality involved in the process of choosing between rival hypotheses or theories that are equally suitable for the available evidence. The overall objective of our thesis is to investigate whether, and to what extent, the acceptance of the underdetermination thesis represents a significant obstacle to scientific realism. Our work will specifically analyze a recent formulation of the problem of underdetermination elaborated by the American philosopher Kyle Stanford in 2006, known as the "Problem of Unconceived Alternatives." We will argue that it is possible to formulate a realistic solution to the challenges arising from the New Pessimistic Induction and the Problem of Unconceived Alternatives. In our thesis, we aim to develop a defense of realism that goes beyond merely discussing the rationality of the belief concerning the existence or absence of unobservable entities. In agreement with Chakravarty (2007), Worral (1989), and Egg (2012), we maintain that a more sophisticated and resilient defense of scientific realism should be based on a realism of "properties and relations," with an emphasis on causal interactions, and should not be restricted to a problematic realism centered around entities.