Grappling with Monsters
How Counter-examples Shape Mathematical Truth
The Euclidean Facade
For too long, mathematics has been presented as a finished cathedral: pristine, silent, and eternal. We are taught the theorems and the proofs as if they descended from the heavens on stone tablets, perfect and immutable. But anyone who has actually done mathematics—who has wrestled with a conjecture in the early morning hours—knows this is a lie.
I have long been influenced by Imre Lakatos and his seminal work, Proofs and Refutations: The Logic of Mathematical Discovery. Lakatos, a student of Sir Karl Popper, argued that the “deductivist” style (definition => theorem => proof) hides the struggle. It hides the history. Most importantly, and the topic of this post, it hides the monsters.
Lakatos’ motivation was radical for his time: he wanted to apply a Popperian philosophy to mathematics. Just as science proceeds by falsification, Lakatos argued that mathematics proceeds by “heuristic falsification.” It is a dynamic, living process where truths are not just discovered, but forged in the fires of criticism.1
Against “Authoritarian” Mathematics
Lakatos famously despised what he called the “formalist”2 approach, which he felt reduced mathematics to meaningless manipulation of symbols. He felt that reducing mathematics to the static manipulation of symbols was a form of intellectual authoritarianism. If you present a proof as a perfect, unbroken chain of logic without showing the struggle, the false starts, and the counter-examples that forced the definitions to evolve, you are hiding the actual work of mathematics.
His goal was to restore the “human face” of mathematics, to reveal the thinking that gives rise to the mathematics. He structured his book, Proofs and Refutations,3 as a classic Socratic dialogue between a teacher and students (named Alpha, Beta, Gamma, etc.). This Socratic setting allowed him to dramatize the birth of a theorem.
Crucially, Lakatos argued for a “quasi-empiricism.” This does not mean he doubted for a moment that 1+1=2 follows from the Peano axioms. In fact, he did not question the validity of deduction within a closed system. Rather, he doubted such closed systems could ever be considered as “final” or fully capture the dynamic, growing body of informal mathematics.
For Lakatos, a formal proof is but a “snapshot” of a living process. The certainty we gain in a formal system is real, but it is purchased at the cost of freezing the concepts in place. In the messy world of discovery, definitions are not fixed; they are negotiated in the face of “monsters.”
Enter the Monsters
The heart of Lakatos’ analysis is the “mathematical monster.” A monster is a counter-example that threatens to destroy, serving as a refutation of established definitions or cherished propositions and proofs.
Lakatos uses Euler’s polyhedral formula as his central case study. Euler stated that for any polyhedron, the number of vertices (V), minus the edges (E), plus the faces (F) equals 2: V - E + F = 2.
Nominally, it seems elegant and obviously true. But then, the monsters arrive.
What about a cube with a hole in it (like a picture frame)?
What about a “crested cube” (a smaller cube sitting on a larger one)?
What about a cylinder (which has 0 vertices and 0 edges)?
When these monsters appear, the mathematician is faced with a crisis of sorts. Lakatos identifies a number of standard defensive tactics we use to protect our theorems, tactics that reveal the sociology of logic.
1. Monster-Barring
This is the most common defence. The mathematician redefines the terms to exclude the monster.
The Monster: “Here is a hollow cube where V-E+F ≠ 2.”
The Response: “That is not a real polyhedron. A polyhedron must be a solid without holes.”
By narrowing the definition, we save the theorem, but we reduce its domain. As Lakatos wryly put it, monster-barring is the method of defining away the counter-example.
2. Monster-Adjustment
In this more subtle tactic, the mathematician reinterprets the monster so that it fits the rule.
The Response: “If we view the cylinder not as a smooth surface, but as a prism with infinite sides, then the formula still holds!”
This is a linguistic sleight of hand—we change our perception of the object to save the integrity of the equation.
The Logic of Discovery and Embracing the Mess
Why does this matter? Because, as Lakatos shows, the refutation is the most valuable part of the process. The monsters are not nuisances; they are the drivers of discovery.
When we encounter a monster, we are forced to move from “naive conjecture” to “proof-generated concepts.” The struggle with the hollow cube forces us to invent the concept of topological genus. The struggle with the crested cube forces us to be precise about what constitutes a “face” or an “edge.”
The theorem grows smarter because it was attacked.
In our modern era of formalized, computer-checked proofs, Lakatos is a necessary corrective. He reminds us that rigour does not mean sterility. The “logic of discovery” is a logic of uncertainty, of negotiation, and of creative re-definition.
We should not fear the mathematical monsters. We should invite them in. It is only by grappling with them that we truly understand the numbers and shapes we study. As I continue my own work in mathematics, I try to remember: a proof is not a wall to keep errors out; it is a path cut through a jungle of monsters.
Lakatos also sought to reconcile Popper’s falsificationist view of the scientific method with Thomas Kuhn’s historical and sociological insights into scientific practice (https://stephenrcampbell.substack.com/p/philosophical-frameworks-and-research).
Formalism largely refers to the "Hilbert Program" of the early 20th century. David Hilbert sought to secure the foundations of mathematics by axiomatizing it completely, that is, basically treating mathematics as a game of manipulating symbols according to strict rules, independent of their semantic "meaning." While Hilbert aimed for absolute certainty (consistency and completeness), Lakatos argued that this "deductivist" style whitewashed the actual history of mathematics, presenting it as a static, perfect structure, wrongfully excluding it as also being a dynamic and evolving process of trial and error.
Whoever borrowed my dog-eared copy of Proofs and Refutations, with three readings worth of marginalia therein, please contact me. I’d love to have that part of my memory back.