[Paper Review] Desperately seeking mathematical truth

Melvyn B. Nathanson|arXiv (Cornell University)|Sep 8, 2008

Philosophy and Theoretical Science19 citations

TL;DR

This paper critiques the perceived certainty of mathematical truth, arguing that many celebrated theorems—such as those by Wiles and Perelman—lack fully verified, line-by-line proofs. Instead, their acceptance rests on political consensus among mathematical elites, revealing that mathematical literature is fundamentally unreliable despite its claim to logical rigor.

ABSTRACT

This article discusses epistemological problems in the philosophy of mathematics and issues concerning the reliability of the mathematical literature.

Motivation & Objective

• To challenge the widespread belief that mathematical theorems are universally and eternally true through rigorous, checkable proofs.
• To expose the systemic flaws in mathematical peer review and the reliance on expert opinion rather than line-by-line verification.
• To illustrate how even landmark proofs—like those of Fermat’s Last Theorem and the Poincaré Conjecture—lack complete verification and depend on community consensus.
• To argue that the mathematical literature is unreliable due to incomplete proofs, unverified references, and flawed refereeing practices.
• To highlight the political nature of mathematical truth, where 'bosses' in fields determine correctness, not logical completeness.

Proposed method

• Analyzes historical and contemporary mathematical proofs, including Wiles’ proof of Fermat’s Last Theorem and Perelman’s proof of the Poincaré Conjecture, to demonstrate reliance on consensus over verification.
• Examines the classification of finite simple groups as a case study of a theorem with no universally accepted complete proof, despite decades of work.
• Reviews the refereeing process in mathematics journals, arguing that most papers are not thoroughly checked and that referees often only skim proofs.
• Contrasts the ideal of mathematical proof—logical deduction from axioms—with the reality of incomplete, sketchy, or intuitive arguments that are accepted as 'true'.
• Uses the historical development of calculus and Euclid’s Elements to show that even foundational mathematics contains logical gaps that were only resolved long after initial publication.
• Argues that the truth of a theorem depends not on its logical completeness but on the social validation by a small group of experts, making mathematical truth a political construct.

Experimental results

Research questions

• RQ1 Why do mathematicians accept major theorems as true when they lack fully verified, line-by-line proofs?
• RQ2 To what extent is the mathematical literature actually reliable, given the prevalence of incomplete proofs and unverified references?
• RQ3 How does the refereeing process in mathematics journals fail to ensure the correctness of published results?
• RQ4 Why do landmark proofs like Wiles’ and Perelman’s take years to be accepted, despite their significance?
• RQ5 In what ways does the concept of mathematical truth depend on social consensus rather than logical completeness?

Key findings

• Many celebrated mathematical theorems, including those by Wiles and Perelman, are not fully verified; their correctness rests on expert consensus, not exhaustive proof-checking.
• The classification of finite simple groups remains controversial, with no universally accepted complete proof, and many results in the literature are still considered unproven assertions.
• Most papers in refereed mathematics journals are not actually refereed in depth; referees often only check statements and skim proofs, leading to widespread acceptance of flawed or incomplete arguments.
• Even foundational works like Euclid’s Elements contain logical gaps that were not corrected for over a millennium, demonstrating long-standing flaws in mathematical reasoning.
• The history of calculus shows that major mathematical fields were used and developed for 150 years before their foundations were rigorously established, indicating that truth in mathematics often follows rather than precedes proof.
• Mathematical truth is not purely logical but political: the judgment of correctness is determined by influential experts or 'bosses' in a field, not by objective verification.