Abstract

In this paper I argue that the values supported by scientists can have a role in episodes of theory choice. In the first part, I characterize the value- and the rulebased accounts of theory choice. In the second part, I analyze how the thesis of underdetermination of theory by empirical data can be used to argue for a value-based account. I discuss two versions of the underdetermination thesis, arguing that the weaker version, underdetermination by the evidence available at a particular time, is sufficient for establishing the role of values in theory selection. Many authors distinguish between cognitive and non-cognitive values, considering that only the former ones have a legitimate role in theory choice. I defend this distinction, showing that it has both a normative and a descriptive dimension. I argue that cognitive values must not be seen as indicators of truth, but they can be characterized by their relation to the goal of science. In the end, I argue that, in spite of being justified and useful, the distinction between cognitive and non-cognitive values is not clear-cut