DS-2015-02: The Kid, the Clerk, and the Gambler - Critical Studies in Statistics and Cognitive Science

DS-2015-02: Madsen, Mathias Winther (2015) The Kid, the Clerk, and the Gambler - Critical Studies in Statistics and Cognitive Science. Doctoral thesis, University of Amsterdam.

	Text (Full Text) DS-2015-02.text.pdf Download (5MB)
	Text (Samenvatting) DS-2015-02.samenvatting.txt Download (2kB)

Abstract

In this dissertation, I have presented a series of case studies in linguistics, psychology, and statistics. These case studies have taken up a variety of theories, concepts, and debates, and have in each case attempted to shed new light on these topics by consistently focusing on foundational issues. What these foundational issues were has depended on the specific topic at hand.
In my discussion of cognitive metaphor theory in Chapters 2 and 3, I thus took issue with the strong emphasis on direct, bodily experience and attempted to show that this physicalistic origin of linguistic behavior should be seen as one among many sources, and probably not the most important one. In my discussion of the syntax-semantics distinction in Chapter 4, on the other hand, I criticized an overly logical "filing cabinet" theory of comprehension and argued that much of the EEG evidence supporting it could be explained more elegantly in terms of probabilistic reasoning and noisy-channel coding.
The next series of case studies combined various conceptual frameworks that are usually seen as separate. In Chapter 5, I thus defined a quantitative concept of relevance based on classical ideas from information theory and decision theory, and I hinted at ways it could be applied to situations in computer science and microeconomics. In Chapters 6 and 7, I selectively transplanted components from epistemic logic into a Bayesian reasoning system in order to construct a flexible modeling language which might have applications in cognitive science and artificial intelligence.
The last two case studies focused on issues regarding the philosophy of statistics and the limits of rationality. In Chapter 8, I discussed the bias-variance trade-off, a conventional way of thinking about the "loading" of statistical estimators, and I explained why this trade-off had no coherent interpretation in Bayesian statistics, even though it seems, from the outside, to support certain tenets of that methodology. In Chapter 9, sketched a history of the protracted war between frequentist and Bayesian statistics and reinterpreted this debate as a conflict between two norms of rationality expressed in mathematical form.

Keywords:

Item Type:	Thesis (Doctoral)
Report Nr:	DS-2015-02
Series Name:	ILLC Dissertation (DS) Series
Year:	2015
Subjects:	Language Logic Philosophy
Depositing User:	Dr Marco Vervoort
Date Deposited:	14 Jun 2022 15:17
Last Modified:	14 Jun 2022 15:17
URI:	https://eprints.illc.uva.nl/id/eprint/2127

Actions (login required)

View Item