Higher-order evidence, reliability, and peer disagreement

  • Nemo A D'Qrill

Student thesis: Doctoral ThesisDoctor of Philosophy (PhD)

Abstract

A minister commissions a report on whether p. A dozen experts submit their credences in whether p but at least two of them assign different credal values. The reliabilities of each expert are known to the minister. Given all these pieces of information, how should the minister, knowing nothing else about p, assign her credence? This question, Δ, is clearly both complicated and important. This thesis provides a novel answer to Δ. We examine various research programs dealing directly or indirectly with Δ under different formal interpretations of what is meant by the experts’ reliability. We show how each program reveals uniquely relevant features but are limited in scope due to formal constraints. We expand upon each program individually and form a more general account in answering Δ. In the process we tidy up terminology, formalise previously casually used concepts, and bring to light key variables such as sensitivity to the experts’ type-I/-II error, previously not given due attention in the literature. Beyond providing a more sophisticated and general answer to Δ, the tidying up of terminology and formalism allows the academics of the different research programs to collaborate and share results when tackling future problems on how to deal with sophisticated types of (higher-order) evidence.
Date of Award27 Sept 2022
Original languageEnglish
Awarding Institution
  • University of Bristol
SupervisorRichard G Pettigrew (Supervisor) & Jason P Konek (Supervisor)

Cite this

'