Dissertation/Thesis Abstract

Formal proofs and refutations
by Alama, Jesse, Ph.D., Stanford University, 2009, 254; 3364492
Abstract (Summary)

Two questions drive the dissertation: (1)What can one discover in a formal mathematical theory? (2) What more do we know of a mathematical theorem when it has been formally proved than that it is provable? These questions spring from the provocative philosophy of mathematics of Imre Lakatos. They are tackled in two ways: by evaluating the philosophical foundations of Lakatos's work, and by studying contemporary work in formal mathematics, specifically formal proof checking technology.

The dissertation has a technical part and a philosophical part. The first part considers some philosophical problems raised (or brought into focus) by formal mathematical proofs. The second, technical part attempts to answer mathematical questions raised in the first part. The bridge between the two is a formal proof of a famous mathematical result known as Euler's polyhedron formula, whose history Lakatos has reconstructed and which serves as the central example for his philosophy of mathematics. The aim of the dissertation is to explore some of the philosophical problems suggested by such formalization efforts.

The argument of the dissertation has three components. In the first component, I explain how Lakatos's philosophy of mathematics poses a challenge to formal proof checking technology. The second component is to respond to the challenge by formalizing Euler's polyhedron formula. Finally, the third component evaluates the technical, formal proof response.

The dissertation is timely because it has now become possible to actually formalize significant mathematical proofs. These contemporary developments are a source of problems for a philosophy of mathematics that is sensitive to mathematical practice. This movement within the philosophy of mathematics is to no small degree inspired by Lakatos's work. The time is ripe for returning to some of the basic philosophical problems that Lakatos and other philosophers pointed to long ago, and to examine new problems that come from the development of what might be called proof technology, tools for helping mathematicians construct and evaluate proofs.

In chapter 1, I lay out some of the main questions and problems about formal proofs and explain how they are related to central issues within mainstream philosophy, particularly epistemology and philosophy of science. The development of formal proof technology is based on classical 19th and 20th century results in mathematical logic but depends crucially on computers. Chapter 1 also surveys the variety of uses of computers in mathematical practice and discusses the variety of philosophical problems they pose.

The next step in the discussion of formal proofs will be a critical evaluation of the philosophy of mathematics of Imre Lakatos. His Proofs and Refutations (1963) attacks formalist philosophies of mathematics. Since much proof technology is to some extent based on or requires a certain formalist view of mathematics, the question naturally arises how Lakatos's philosophy bears on these developments. Chapter 2 addresses these concerns. I focus also on some epistemological problems suggested by formal proofs, such as the question of defining rigor and the problem of whether and how one improves one's justification for a mathematical claim through formalization of a proof of it.

The cornerstone of Lakatos's Proofs and Refutations is a history of a particular mathematical theorem known as Euler's polyhedron formula, which is a certain geometrical-combinatorial claim with a rather colorful history. I have formalized a proof of this mathematical result; chapter 3 contains a discussion of the proof and its formalization.

Thanks to the work carried out in chapter 3, Euler's polyhedron formula (understood in a certain abstract or combinatorial way that is explained in chapter 3) is shown to be a first-order consequence a certain first-order theory of sets. Because of the peculiarities of the particular proof technology with which the formal proof was carried out, the theory of sets that is used is much stronger than what is intuitively required for Euler's theorem. A natural proof-theoretic question thereby arises: can one do better? Are the strong assumptions really necessary? In chapter 4, I carry out the project of trying to identify a weaker theory in which to carry out a formal proof Euler's formula. Also discussed are some formal problems about expressibility problems for combinatorial polyhedra, and related issues.

In chapter 5, I return to some of the issues that Lakatos raised in connection with formal proofs in light of the formal work that is carried out in chapter 3. This work provides some resources for taking on the two questions that were initially asked. I show that Lakatos's philosophy, its strong reservations against 'formalism' notwithstanding, applies quite naturally to formal mathematics. (Abstract shortened by UMI.)

Indexing (document details)
Advisor: Mints, Grigori
School: Stanford University
School Location: United States -- California
Source: DAI-A 70/07, Dissertation Abstracts International
Subjects: Philosophy
Keywords: Automated reasoning, Epistemology, Formal proofs, Lakatos, Imre, Philosophy of mathematics, Proof technology
Publication Number: 3364492
ISBN: 9781109242775
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy