15th Graduate Conference in Political Philosophy – Abstract/Costa

Luca Costa
The Puzzle of Evidence: How to Ground Normative Claims in Empirical Data

The usage of empirical data to support the claims of normative political theories is widespread in political philosophy. However, the methodology which grounds this usage is rarely discussed by political philosophers, beyond questions concerning the general admissibility of empirical facts as premises to normative arguments. In this paper, I focus on the issues raised by this methodology, with particular attention to the contributions from the social sciences, such as economics, psychology and anthropology.
I begin with an overview of what I call the piecemeal approach to empirical evidence, which characterizes the majority of works in normative political philosophy. According to this approach, empirical studies are to be treated by a normative political theory as discrete pieces of evidence. The theoretical frameworks these studies are based on, their underlying empirical assumptions, and the consistency of these assumptions need not be discussed within the normative theory itself.
Then, I proceed to examine the limits of this piecemeal approach. In particular, I argue that this approach fails to account for the different, and often conflicting, empirical assumptions which ground the aforementioned social sciences. I also analyse Gaus’s theory of public reason (Gaus 2011), whose appeal to extensive empirical evidence further illustrates the limits of the piecemeal approach in weighing the data produced by the social sciences.
Finally, I suggest an alternative approach to the use of empirical data in a normative political theory, which organizes the data – as well as the empirical assumptions of the social sciences which produce and interpret these data – in a consistent empirical account of human nature and society. I conclude arguing how a normative political theory can be based on such an empirical account, while respecting the logical priority of normative principles over facts (Cohen 2003).