ON PROOF AND PROGRESS IN MATHEMATICS

✨ Summary

Thurston argues that the primary accomplishment of mathematics is not the production of definitions, theorems, and proofs, but the advancement of human understanding. He challenges the definition–theorem–proof model because it does not explain where mathematical questions come from or how mathematical insight is developed.

The paper presents understanding as multidimensional. Mathematicians use symbolic, logical, geometric, linguistic, intuitive, metaphorical, procedural, and computational modes of thought. These modes are not interchangeable: a formally correct statement may fail to communicate the mental models that make a result meaningful. Thurston therefore treats mathematical communication as a central component of mathematical progress rather than as a secondary activity.

The paper also describes mathematical knowledge as partly social and community-based. Within a specialist field, knowledge is transmitted efficiently through shared background, informal conversation, seminars, and personal interaction. Formal papers provide durable records, but they often omit the informal infrastructure through which mathematicians actually understand results. Proof is consequently portrayed as a humanly checkable argument embedded in a social process of criticism, explanation, trust, and reuse. Thurston does not advocate weaker standards of rigor; instead, he recommends proofs that are clearer and simpler so that possible weaknesses are easier to detect.

Computers provide a contrast: they are well suited to formal verification but poor at supplying human understanding. Thurston supports computer formalization as a valuable long-term project, while distinguishing machine-checkable formal deductions from the kinds of proofs mathematicians normally use and understand.

Thurston further argues that mathematical progress depends on many contributors, not only on the individuals who receive credit for proving major theorems. Developing terminology, examples, conceptual infrastructure, explanations, alternative proofs, teaching materials, and research communities can be as important to the growth of a subject as obtaining the initial result. He criticizes academic incentives that place excessive emphasis on theorem-credit.

His experiences with foliations and three-manifold geometry illustrate these claims. Results in foliation theory were technically significant but difficult to enter because the surrounding conceptual infrastructure was insufficiently communicated. In contrast, his later work on geometrization was accompanied by extensive efforts to develop shared language, examples, intuition, notes, seminars, and cross-disciplinary communication. He concludes that creating an environment in which others can understand, extend, and apply new ideas may have greater long-term value than rapidly publishing every theorem.

The essay has subsequently been used as a reference in scholarship on mathematical proof, mathematical communication, pedagogy, and the philosophy of mathematical practice. Later educational and philosophical works cite it when discussing the role of multiple representations, explanation, social validation, and human understanding in mathematical proof and learning. (arxiv.org)