Number Theory Theorems Whose First Proofs Were Long or Difficult
At a glance
- 1. Define the comparison
Proof length and proof difficulty are related but different; judge each by its historical context.
- 2. Compare examples
Look at landmark results in prime distribution, Diophantine equations, and arithmetic geometry.
- 3. Check the history
Separate the earliest accepted proof from later repairs, simplifications, or clearer expositions.
- 4. Read with context
A difficult proof may be short on paper but rely on substantial new theory.
“Long” and “difficult” are not the same thing. A proof may take many pages because it handles numerous cases, or it may be concise but depend on powerful machinery that took decades to develop. Historical comparisons also require care: the first published proof, the first complete proof, and the first proof most readers can follow are not always the same.
Here are several well-known examples from number theory. They are illustrations, not a definitive ranking.
Dirichlet’s theorem on primes in arithmetic progressions
Dirichlet’s theorem says that if positive integers and have no common factor other than , then there are infinitely many primes of the form , where is a nonnegative integer. For instance, there are infinitely many primes congruent to modulo , and infinitely many congruent to modulo .
Dirichlet proved the theorem in 1837. The argument introduced characters and used analytic tools, including a key nonvanishing result for certain -series at . The proof was a major achievement partly because the statement is about primes in each individual admissible residue class, not merely about primes in general. Later textbooks often present a more systematic version, but the central ideas remain substantial.
The prime number theorem
The prime number theorem describes the overall frequency of primes. If denotes the number of primes at most , it states that
as grows without bound. The symbol means that the ratio of the two sides tends to .
The first proofs were published independently in 1896 by Jacques Hadamard and Charles-Jean de la Vallée Poussin. Both used complex analysis to study the Riemann zeta function and establish that it has no zeros on a crucial part of the line with real part . The proof is demanding not simply because it is long, but because it connects prime counting to analytic properties of a complex function. Later proofs have used different methods, but the original result marked a major turning point in analytic number theory.
Faltings’ proof of the Mordell conjecture
The Mordell conjecture, now called Faltings’ theorem, says that a smooth algebraic curve of genus greater than has only finitely many rational points. A rational point is a solution whose coordinates are rational numbers, subject to the equations defining the curve.
Gerd Faltings proved the conjecture in 1983 using ideas from arithmetic geometry, including results about abelian varieties and heights. The proof was a major theoretical advance, not a short elementary argument. Its significance is also visible in the response to it: later work found new perspectives and consequences, but the original proof established a general finiteness theorem that had resisted attack for decades.
Fermat’s Last Theorem
Fermat’s Last Theorem says that for every integer , there are no positive integers satisfying
Andrew Wiles announced a proof in 1993, but a gap was found in one part of the argument. Wiles and Richard Taylor then developed a repair, and the corrected proof appeared in 1995 in two papers. The work links elliptic curves and modular forms; it is not a direct attack on the equation by elementary manipulations. The published proof is technically advanced and lengthy, and its ideas grew out of a much broader program in modern number theory.
This is also a useful reminder about what “first proof” can mean. Wiles’s announced argument was not the final accepted proof; the corrected work is the proof generally credited with establishing the theorem.
A useful way to compare these examples
| Result | Main source of difficulty |
|---|---|
| Dirichlet’s theorem | Character sums and nonvanishing of -series |
| Prime number theorem | Complex analysis and zeros of the zeta function |
| Faltings’ theorem | Arithmetic geometry and abelian varieties |
| Fermat’s Last Theorem | Deep links between elliptic curves and modular forms |
These examples show why page count alone is a poor measure. A proof can be long because it spells out a complicated argument, but it can also be difficult because a short statement relies on an extensive technical framework. Conversely, a later proof may be shorter because it can use concepts and theorems unavailable to the original authors.
Common mistakes when discussing first proofs
- Treating difficulty as an objective ranking. What seems hard depends on the reader’s background and on what tools are allowed.
- Confusing a proof’s length with its depth. A compact argument using advanced theory may be harder to understand than a longer elementary proof.
- Ignoring repairs and revisions. An announced proof with a gap is not the same as a completed, accepted proof.
- Calling an old proof elementary by modern standards. A result may now have a short proof because later mathematics supplied new tools; that does not make its first proof easy.
For comparisons with results whose solutions were unexpectedly short, see discussions of short proofs of open problems and short but substantial mathematics papers.
References
- H. Davenport, Multiplicative Number Theory
- G. Faltings, “Endlichkeitssätze für abelsche Varietäten über Zahlkörpern,” Inventiones Mathematicae 73 (1983), 349–366
- Andrew Wiles, “Modular elliptic curves and Fermat’s Last Theorem,” Annals of Mathematics 141 (1995)
Spotted an error? Send a correction. Read how we write and check articles in our editorial policy.