Dissertation/Thesis Abstract

Skill-based Reliabilism
by Marshall, Daniel C., Ph.D., University of Missouri - Columbia, 2014, 273; 13877155
Abstract (Summary)

Reliabilism, at its most general, is the claim that a belief has some positive epistemic status (typically justification or knowledge) only if it was formed by a reliable belief-forming process or mechanism. One important problem facing reliabilism is the Generality Problem. The Generality Problem arises because reliability is generally taken to be a property of a type of mechanism or process, and a given token process belongs to as many different types as it has properties. The Generality Problem consists of the task of specifying in a principled manner which of the multitude of types a given belief-forming token-process belongs to is the one whose reliability is relevant to the epistemic project.

In this dissertation, I will propose a solution to the Generality Problem. My proposed solution is based on a model of skill-learning which I have developed based on the theories of motor-learning provided by John A. Adams and Richard A. Schmidt, and a theory of cognitive architecture (which focuses on cognitive skills) provided by John R. Anderson. This model is most directly applicable to cases where beliefs are formed by acquired cognitive skills, but it can easily be extended to apply to cases involving innate cognitive instincts, and, with some degree of difficulty, to cases involving belief formation through a process of deliberation.

I will develop my proposed solution to the Generality Problem in two stages. In the first stage, I will propose a “rough draft” version of the solution. According to this rough draft version, the mechanism type that is relevant to the epistemic project is a mechanism’s activation type. A belief-forming mechanism of a given activation type consists of an activation mechanism which is responsible for activating ‘a’ belief-forming program that is stored at a certain address. This belief-forming program is, at least typically, acted on by learning, selection, and/or maintenance mechanisms, which constrain the degree and manner of how the program can vary from time-slice to time-slice and from world to world. Any belief-forming program (or version of the same program) that is activated by the same activation mechanism belongs to the same activation type.

This rough draft solution is incomplete in two ways. First, it provides no way of specifying the epistemically relevant type a given activation mechanism belongs to. Secondly, it provides no obvious reason to think that this type, out of the multitude that a given belief-forming process-token or mechanism-token belong to, is relevant to the epistemic project.

These two issues are addressed by a more refined version of my proposed solution, which claims that the type a belief-forming mechanism belongs to which is relevant to the epistemic project is its extended activation type. This type is also identified on the basis of the activation mechanism, but this proposed solution recognizes that the epistemically relevant type of the activation mechanism itself can be specified by some underlying adaptive mechanism (a learning, selection, developmental, maintenance, etc. mechanism). This mechanism’s type, in turn, is specified by some further underlying adaptive mechanism. This regress can be non-problematically terminated by such adaptive mechanisms as evolution by natural selection or a certain kind of Intelligent Design. The reliability of this type is plausibly linked to the epistemic project, since it incorporates the entire flow of information that went into shaping the belief-forming skill, including information that is not reflectively accessible by the subject.

This proposed solution provides some traction for solving two other problems that confront reliabilism: the New Evil Demon Problem and the Strange and Fleeting Process Problem. In the course of this dissertation, I will use my proposed solution to the Generality Problem to provide solutions to two versions of the New Evil Demon Problem, and to provide partial solutions to cases involving a third version of the New Evil Demon Problem. I will also use my proposed solution to the Generality Problem to provide solutions to several Strange and Fleeting Process problem cases. The basic strategy I will use to solve these Strange and Fleeting Process problem cases is that the belief-forming processes involved in these cases tend to belong to a ‘common-sense’ type which is reliable, but to an extended activation type that is unreliable.

Indexing (document details)
Advisor: Weirich, Paul
Commitee: Ariew, Andre, Markie, Peter, Smith, George, Weirich, Paul
School: University of Missouri - Columbia
Department: Philosophy
School Location: United States -- Missouri
Source: DAI-A 80/08(E), Dissertation Abstracts International
Subjects: Philosophy
Keywords: Epistemology, Generality problem, Proper function, Reliabilism, Skill
Publication Number: 13877155
ISBN: 9781392055748
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy