12th Pavia Graduate Conference in Political Philosophy – Abstract/Favara


Greta Favara

Abstraction and Idealization: An Irrelevant and Misguiding Methodological Divide

Despite its wide acceptance and its seemingly uncontroversial character, the methodological divide between abstraction and idealization in political theory, if carefully examined, reveals itself to be worthless and counterproductive. In this paper, my attempt will be to show why we should overcome this well-established distinction, originally introduced by Onora O’Neill. I claim that there are two fundamental reasons for doing so. Firstly, this methodological divide does not give any substantive advice in order to build a normative political argument – I will call this first rebuttal the irrelevancy objection. Secondly, and more problematically, the divide contributes to obscure some crucial methodological issues that should be directly faced and clarified – I will call this second challenge the misguidance objection. Both objections are grounded on the idea that O’Neill endorses a naive and insufficient conception of description. Indeed, the distinction between abstraction and idealization refers to some features of the description adopted: it depends exclusively on its truthfulness or falsehood. Specifically, abstractions are subtractions of true predicates, idealizations are additions of false predicates. According to this distinction, for O’Neill idealization should be avoided, while abstractions are necessary. However, I maintain that focusing merely on the truth or falsehood of premises obscures the relevant methodological relationship, which is triadic and not binary. Indeed, the description a theory endorses is never selected just for its truth-value. The relevant methodological link is not that between theory and (the truth-value of) descriptive assumptions, but runs over a triangulation between theory, descriptive assumptions and criterion of appropriateness. As a consequence I claim that, even if there is a fundamental logical difference between abstraction and idealization, this sharp divide is methodologically unfruitful: focusing solely on such a binary relationship is at best useless, at worst harmful.