Discovering the Impossible — When Failure Became a Theorem

Change the Tool and the Angle Splits into Three — Archimedes' Neusis

A construction attributed to Archimedes trisects a general angle by sliding a ruler carrying a marked length until two conditions meet, a procedure called neusis. The same goal is generally impossible with unmarked straightedge and compass but possible with a marked ruler or special curves. Impossibility belongs to a goal together with its rules.

Translating Geometric Problems into Degrees of Equations — Descartes

La Géométrie, printed in Leiden, advanced a way to treat segment relations algebraically and curves by equations. It opened a path for translating constructibility into properties of numbers and polynomials. Descartes did not anticipate Wantzel's 1837 theorem, but algebra became an essential shoulder for proving geometric limits.

Seeking a Proof That the Formula Does Not Exist — Ruffini

Ruffini published a book claiming that general equations above degree four could not be solved by radicals. His early arguments contained assumptions and gaps that were not accepted, and contemporary response was limited. Yet he redirected Lagrange's study of permutations from searching for a formula toward proving a formula's limits, preparing a structural path toward Abel and group theory.

Proving There Is No Radical Formula for the General Quintic — Abel

At twenty-two, Abel paid to print an extremely compressed memoir in Christiania proving that no universal formula using arithmetic and radicals expresses the roots of the general quintic. It does not say quintics have no roots or cannot be approximated numerically. It is an exact impossibility relative to the permitted language of radicals.

Why Are Some Equations Solvable and Others Not? — The Galois Criterion

Patronage and power · lens spotlight

Galois went beyond ruling out one universal quintic formula and opened a way to decide solvability by the structure of relations preserved when roots are permuted. His 1830 Academy submission was not accepted during his life; Liouville published major manuscripts in 1846. Modern group and field terminology should not be projected unchanged backward, but symmetry had turned impossibility into structure.

Turning Two Millennia of Failed Constructions into a Theorem — Wantzel

Wantzel used the condition that straightedge-and-compass lengths arise through successive quadratic extensions to prove that doubling the cube and trisecting a general angle are impossible under those rules. Some angles can be trisected and other tools handle the general case. The theorem proves inaccessibility by specified operations, not merely a long record of failure.

The Transcendence of Pi Makes Squaring the Circle Impossible — Lindemann

When Lindemann proved that pi is transcendental—not a root of any polynomial with rational coefficients—it followed that straightedge and compass cannot construct a square exactly equal in area to a given circle. Physical approximation and other tools remain possible. What is ruled out is an exact finite Euclidean construction.

A Request for a Method Receives a Negative Answer — Hilbert's Tenth Problem

Hilbert's Paris problem list asked for a finite procedure deciding whether an integer-coefficient polynomial equation has an integer solution. The modern definition of algorithm did not yet exist, but the request was for a universal procedure that halts and answers every input. Seventy years later, the complete answer was a proof that no such procedure exists.

Asking for a Procedure That Decides Every Logical Sentence — The Decision Problem

Patronage and power · lens spotlight

Hilbert and Ackermann's textbook sharply asked whether a finite mechanical procedure can decide if a first-order sentence is valid in every interpretation. Göttingen's lectures, textbooks, and proof-theory community shared a precise goal before computation itself was fully defined. Completeness, consistency, and decidability are distinct questions.

Incompleteness Is Announced at a Meeting Expecting Completeness — Gödel

In a discussion at the Königsberg conference of September 1930, Gödel briefly announced his first incompleteness result. In the same city the next day, Hilbert delivered the speech ending “We must know, we will know.” The scenes are not a simple duel: increasingly precise formalization made it possible to state and prove a limit of formal systems.

A System Reads Its Own Sentences and Finds the Boundary of Proof — Incompleteness

Gödel encoded formulas and proofs as natural numbers so arithmetic could speak about its own provability. Under specified conditions, a consistent, effectively axiomatized system strong enough for arithmetic contains statements it neither proves nor disproves. This does not make every mathematical truth forever unknowable or make first-order logic itself incomplete.

Refuting a Universal Decision Procedure with Lambda Calculus — Church

Patronage and power · lens spotlight

In Princeton, Church connected effective calculability with lambda definability and recursive functions and proved the first-order decision problem has no universal solution. Once effective procedure became a mathematical object, its impossibility became provable. Church and Turing followed independent formalisms that were soon connected as equivalent in reach.

Turning the Human Calculator into a Machine and Proving Some Runs Cannot Be Decided — Turing

Turing modeled a person reading squares, writing symbols, and following finitely stated rules as an abstract machine. A diagonal argument showed there is no universal machine deciding whether another machine will continue producing symbols, yielding a negative answer to the decision problem. This is closely related to the modern halting theorem without making every textbook formulation identical to Turing's paper.

There Is No Universal Decider for Integer Equations — The DPRM Theorem

Patronage and power · lens spotlight

Matiyasevich represented exponential growth from Fibonacci numbers with Diophantine equations, resolving the remaining conjecture in Julia Robinson's program. Combined with work by Davis, Putnam, and Robinson, it proved that no algorithm always decides whether an arbitrary integer-coefficient polynomial has an integer solution. Individual equations can still be solved.