SOPhiA 2021

Salzburgiense Concilium Omnibus Philosophis Analyticis

SOPhiA ToolsDE-pageEN-page

Programme - Talk

The Dutch Book Argument for Probabilism and the Expected Utility Objection
(Epistemology, )

A This paper aims to evaluate a novel objection against the Dutch Book Argument for Probabilism: the expected utility objection that has been first put forward by Hedden (2013) and further discussed by Wronski & Godziszewski (2017). The Dutch Book argument (DBA) assumes the so-called Ramsey's Thesis (RT), according to which there is a particular kind of connection between an agent?s credences and her fair betting quotients. Namely, on RT, an agent's credence in a proposition matches her fair betting quotient for that proposition. And according to the expected utility objection, RT is false because it conflicts with the principle of maximising expected utility. Hence, the DBA is unsound; or so the objection goes. This paper argues that the new objection is not successful. My argument, briefly, is as follows: RT amounts to a betting interpretation or explication of an imprecise concept of credence. More fully, on this betting interpretation, the proposition "an agent's degree of belief in H should be q" or explicated as follows: "an agent is indifferent between buying or selling the bet for Sq." q is called an agent's betting quotient for H. The proponents of the expected utility objection, such as Hedden, do not provide any interpretation or explication of credence. Instead, Hedden (and Wronski and Godziszewski) take the concept of credence as primitive or uninterpreted. Now, I agree with Hedden that if we take an uninterpreted, primitive concept of credence, then, given some additional assumptions, an agent's degree of belief may not match her fair betting quotients. But this does not show that the expected utility objection against the DBA succeeds. This is because the objection does not address the DBA as an argument by interpretation. And Hedden and other supporters of this objection do not provide any alternative interpretation or explication of credences. I conclude that an unanalysed notion of credence, assumed by the proponents of the expected utility objection, is inferior to the betting interpretation. Thus, the criticism of RT is unfounded.

Chair: Kimon Sourlas-Kotzamanis
Time: 10:00-10:30, 10 September 2021 (Friday)
Location: HS E.002
Remark: (Online Talk)

Nina Abesadze 
(University of Salzburg, )

Testability and Meaning deco