Regular Primes

For centuries, Fermat's Last Theorem stood as one of mathematics' most famous unsolved problems, a deceptively simple statement that resisted all attempts at proof. A promising 19th-century strategy involved dissecting the equation in richer, more complex number systems. However, this promising path led to a shocking discovery: in these new worlds, the fundamental laws of arithmetic, specifically the unique way numbers factor into primes, seemed to break down completely. This breakdown threatened to derail the entire effort, creating a profound gap in mathematical understanding.

This article explores the concept of ​​regular primes​​, Ernst Kummer's ingenious solution to navigate this crisis. We will see how this idea not only provided a partial proof for Fermat's Last Theorem but also laid the groundwork for entirely new fields of modern number theory. The first chapter, ​​"Principles and Mechanisms"​​, will uncover the crisis of unique factorization, introduce the theory of ideals, and reveal the startling connection between the regularity of primes and the seemingly unrelated Bernoulli numbers. We will then trace this idea's remarkable legacy in the chapter on ​​"Applications and Interdisciplinary Connections"​​, following its thread from Wiles's ultimate proof of Fermat's theorem to the heart of Iwasawa theory and beyond.

Imagine you are a master watchmaker. You have a deep, intuitive understanding of how gears and springs work in the familiar world of brass and steel. Now, someone hands you a watch made of a strange, crystalline material. You open it up, and to your horror, the gears seem to jump and skip. A gear that should turn once might suddenly represent three turns, or none at all. The basic, reliable clockwork you built your career on has failed. This was the situation faced by 19th-century mathematicians, chief among them Ernst Kummer, as they tried to solve one of history's most famous problems: Fermat's Last Theorem.

The theorem states that for any integer exponent $p$ greater than 2, the equation $x^p + y^p = z^p$ has no solutions in positive integers. The approach was to move the problem into a richer world of numbers, but in doing so, they found the very laws of arithmetic seemed to change beneath their feet.

In our everyday world of whole numbers, we have a bedrock principle, so fundamental we often forget it's there: the ​​Fundamental Theorem of Arithmetic​​. It guarantees that any integer can be broken down into a product of prime numbers in exactly one way (ignoring the order). The number 12 is $2 \times 2 \times 3$, and that's the end of the story. This property is called ​​unique factorization​​.

To attack Fermat's Last Theorem, mathematicians factored the equation $x^p + y^p = z^p$ in a new number system. Instead of just integers, they used the ​​cyclotomic integers​​, numbers of the form $a_0 + a_1\zeta_p + a_2\zeta_p^2 + \dots + a_{p-2}\zeta_p^{p-2}$, where the $a_i$ are integers and $\zeta_p$ is a "primitive" $p$-th root of unity (think of it as a point on the unit circle in the complex plane such that $\zeta_p^p = 1$ but no smaller power is 1). The factorization looks beautiful:

$z^p = x^p + y^p = (x+y)(x+y\zeta_p)(x+y\zeta_p^2)\cdots(x+y\zeta_p^{p-1})$

The dream was simple: if this new world of numbers, which we denote $\mathbb{Z}[\zeta_p]$, also had unique factorization, the proof would almost write itself. We have a product of $p$ factors on the right equalling a $p$-th power on the left. If these factors are "coprime" (sharing no common divisors), then each factor must itself be a $p$-th power, up to some trivial unit factors. This leads to a contradiction, proving the theorem.

But here lies the wreckage. Kummer discovered that in many of these cyclotomic number systems, unique factorization fails! It's like finding that $6$ could be factored as $2 \times 3$ and also as, say, $A \times B$, where $A$ and $B$ are new "primes" completely unrelated to 2 and 3. The clockwork was broken.

Instead of abandoning the new number systems, Kummer performed one of the most brilliant salvage operations in the history of mathematics. He proposed that if the numbers themselves were misbehaving, perhaps there was a deeper level of reality where order was restored. He invented the concept of ​​ideal numbers​​, or what we now call ​​ideals​​.

An ideal is not a single number, but a set of numbers—specifically, a set of all multiples of some number, and sums of such multiples. For example, the set of all even integers forms an ideal. In the world of ideals, unique factorization is always restored. Every ideal can be written as a unique product of "prime ideals." Kummer had found the true, underlying atomic structure of arithmetic.

But this beautiful fix came with a cost. In the familiar integers, every ideal corresponds to a number. The ideal of all even numbers is just all multiples of 2, so we can say it's "generated" by 2. We write this as $(2)$. An ideal generated by a single number is called a ​​principal ideal​​. The problem in the new number systems was that some ideals were not principal. There were "ghost" ideals that couldn't be described by any single number.

The degree of this problem is measured by a finite group called the ​​ideal class group​​, let's call it $\mathrm{Cl}(K)$. Its size, the ​​class number​​ $h_K$, counts how many different "types" of non-principal ideals exist. If $h_K=1$, all ideals are principal, and we have unique factorization of numbers. If $h_K > 1$, we don't. The watch is still broken, but now we have a manual that describes the nature of the break.

For his proof of Fermat's Last Theorem, Kummer realized he didn't need a perfectly functioning watch ($h_K=1$). He just needed to be sure that the specific gear associated with the prime $p$ wasn't broken. He needed to ensure there was no "p-torsion" in the mechanism.

This led to the crucial insight at the heart of our story. The proof could still work as long as the class number $h_K$ was not divisible by the prime $p$. A prime $p$ for which this holds is called a ​​regular prime​​.

Why is this condition "good enough"? If $p$ does not divide $h_K$, it means the class group has no elements of order $p$. In our analogy, no gear gets "stuck" for $p$ rotations before clicking over. This has a magical consequence: if you have an ideal $\mathfrak{a}$ whose $p$-th power, $\mathfrak{a}^p$, is a principal ideal, then the regularity condition forces $\mathfrak{a}$ itself to be a principal ideal.

This was exactly the tool Kummer needed. Back in the FLT factorization, he could show that the ideal generated by each factor, $(x+y\zeta_p^k)$, was the $p$-th power of some other ideal. With regularity, he could leap from "ideal is a $p$-th power of an ideal" to "number is a $p$-th power of a number (times a unit)." The original, simple proof strategy was back on the table, and it worked beautifully for all regular primes.

This is a fantastic theoretical breakthrough. But it leaves a daunting practical problem. How on Earth do you check if a prime is regular? Calculating the class number $h_K$ is monstrously difficult. It would be like having to disassemble the entire universe to check the weather.

Here we arrive at one of the most stunning, almost surreal, connections in all of mathematics. Kummer discovered a "barometer"—a simple test to check for regularity that involves a completely different family of numbers, ones that seem to have nothing to do with cyclotomic fields. These are the ​​Bernoulli numbers​​.

The Bernoulli numbers, denoted $B_k$, are a sequence of rational numbers defined by a simple-looking generating function:

They pop up in calculus, in the formula for sums of powers ($1^k + 2^k + \dots + N^k$), and in special values of the Riemann zeta function. They are workhorses of classical analysis.

Kummer's criterion is this: ​​An odd prime $p$ is irregular if and only if $p$ divides the numerator of one of the Bernoulli numbers $B_2, B_4, \dots, B_{p-3}$​​. A prime that is not irregular is, of course, regular.

This is astounding. To check the intricate algebraic structure of a number field, you just have to compute a list of seemingly unrelated rational numbers and check their numerators for divisibility. The first few primes (3, 5, 7, ..., 31) are all regular. But then we find $p=37$. A computer can churn through the Bernoulli numbers and find that the numerator of $B_{32}$ is divisible by 37. Therefore, 37 is an ​​irregular prime​​. The proof of FLT for exponent 37 required new, more difficult ideas. Another famous example is the prime 691, which dramatically appears in the numerator of $B_{12} = -\frac{691}{2730}$, branding 691 as an irregular prime.

The failure of FLT1 for a prime $p$ would imply that $p$ is irregular, so this criterion became a powerful sieve.

Modern number theory gives us even sharper tools, allowing us to dissect the class group with the precision of a surgeon. The symmetries of the cyclotomic field, governed by its Galois group, act on the class group. A particularly important symmetry is complex conjugation (swapping $\zeta_p$ with its reciprocal $\zeta_p^{-1}$). This allows us to split the $p$-part of the class group into two pieces:

• The ​​plus-part​​ ($\mathrm{Cl}(K)[p^\infty]^+$), which is invariant under complex conjugation.
• The ​​minus-part​​ ($\mathrm{Cl}(K)[p^\infty]^-$), which is inverted by it.

A profound result, the ​​Herbrand-Ribet Theorem​​, tells us that the entire "irregularity" problem—the connection to Bernoulli numbers—lives exclusively in the minus-part. A prime $p$ is irregular if and only if the minus-part of its class group is non-trivial.

We can go even further. The entire class group can be broken down into "eigenspaces," each corresponding to a specific frequency, or character, of the Galois group. The Herbrand-Ribet theorem sharpens to say that the eigenspace corresponding to a certain character $\omega^{1-k}$ is non-trivial if and only if $p$ divides the numerator of the specific Bernoulli number $B_k$. For a regular prime like $p=23$, since it divides none of the relevant Bernoulli numerators, we know instantly that all these corresponding minus-eigenspaces are trivial, each having an order of 1. It's like listening to an orchestra and knowing that because a certain dial is set to zero, an entire section of instruments must be silent.

The minus-part of the class group is now well-understood. Its structure is precisely mirrored by the divisibility properties of Bernoulli numbers. But what about the plus-part? This piece is connected to the class group of the maximal real subfield of our cyclotomic world, $\mathbb{Q}(\zeta_p + \zeta_p^{-1})$.

Here, we stand at the edge of modern research. A long-standing, unproven statement known as ​​Vandiver's Conjecture​​ asserts that the plus-part is always "regular" in the p-sense. That is, the class number of the real subfield, $h(K^+)$, is never divisible by $p$.

This would mean that all the "p-trouble" in the class group, all the irregularity detected by Bernoulli numbers, is confined to the minus-part. The plus-part, in this respect, is always simple. If true, Vandiver's conjecture would be equivalent to saying that the plus-part of the $p$-class group, $\mathrm{Cl}(K)[p^\infty]^+$, is always trivial. While the conjecture has been verified by computers for millions of primes, a general proof remains elusive.

From a 19th-century crisis in arithmetic, to a brilliant fix with ideals, to a magical link with Bernoulli numbers, and finally to a deep structural understanding that still holds unsolved mysteries, the story of regular primes is a perfect illustration of the mathematical journey. It's a tale of broken clockwork transformed into a beautiful, intricate, and still-ticking mystery.

In our last discussion, we met the "regular primes"—those well-behaved numbers that Ernst Kummer had hoped would vanquish Fermat's Last Theorem. We saw that their regularity, a seemingly technical condition involving Bernoulli numbers and class groups, sorted the primes into two camps. You might be left wondering, "So what?" Is this just a historical curiosity, a footnote in a failed attempt at a famous problem? The answer, wonderfully, is a resounding no. The story of regular primes is a spectacular example of how a deep question, even one that leads to a temporary impasse, can blossom into whole new fields of thought. The chase for Fermat's ghost led mathematicians into a strange new world, and the tools they forged there have proven to be more valuable than the original prize. In this chapter, we will follow the tracks of this idea, from the final, triumphant solution to Fermat's puzzle into the very heart of modern number theory and even to some unexpected places beyond.

Kummer’s work was a heroic advance, but the irregular primes were the stubborn holdouts. For over a century, the problem stood. The final solution, when it came from Andrew Wiles, was a masterpiece of modern mathematics. It didn't use regular primes directly in Kummer's fashion, but it was the ultimate fulfillment of his grand strategy: to understand an equation about numbers, translate it into a problem about a more complex, richer mathematical structure.

The strategy was as audacious as it was beautiful. Suppose, just for a moment, that someone handed you a solution to Fermat's equation for a large prime exponent $p$, a set of integers $(a, b, c)$ where $a^p + b^p = c^p$. Wiles and others, building on insights of Gerhard Frey, showed that you could take this triplet of numbers and use it to construct a very particular object: an elliptic curve. This isn't just a curve you can draw on paper; it's a rich algebraic entity with its own universe of properties.

Now, this "Frey curve," born from a hypothetical Fermat solution, would be an odd creature indeed. It would be "semistable," but in a way that left a strange signature in its DNA, a signature tied to the prime $p$. The monumental breakthrough was to connect this world of elliptic curves to another, seemingly unrelated universe: the world of modular forms. The Modularity Theorem (once a conjecture by Taniyama, Shimura, and Weil) acts as a grand dictionary, asserting that essentially every elliptic curve over the rational numbers has a corresponding modular form, its "true love" in another mathematical dimension.

So, if our Frey curve existed, it too must have a modular form partner. But here comes the twist, the dramatic final act. Through a breathtaking chain of logic involving deep results like Ribet's level-lowering theorem, it was shown that this partner couldn't be just any modular form. It would have to be a specific type—a weight 2 newform of "level 2". And here's the punchline: mathematicians had long known that the space of such forms is empty. There are no such objects. Zero. It's like a physicist proving that a new theory demands a particle that weighs -1 kilogram.

The conclusion is inescapable. The chain of reasoning is flawless, so the only thing that can be false is the initial assumption that started the whole cascade: the existence of a solution to Fermat's equation. The ghost was laid to rest, not by a head-on assault, but by showing that its existence would violate a fundamental truth about the fabric of the mathematical cosmos. Kummer’s idea of using associated structures had found its ultimate expression.

The journey sparked by regular primes didn't end with Fermat. In fact, a much grander story was just beginning. The class group, that measure of complexity for cyclotomic fields, became an object of intense study in its own right. A new question arose: what happens if we don't just look at the field of $p$-th roots of unity, $\mathbb{Q}(\zeta_p)$, but at an entire infinite tower of fields: $\mathbb{Q}(\zeta_p)$, then $\mathbb{Q}(\zeta_{p^2})$, then $\mathbb{Q}(\zeta_{p^3})$, and so on, climbing up to infinity?

This is the domain of Iwasawa theory. It's like moving from studying a single photograph to watching a full movie. Iwasawa theory provides the tools to understand how arithmetic properties, like the size of the $p$-part of the class group, evolve as you ascend this tower. Miraculously, this growth isn't chaotic. It follows a stunningly simple asymptotic formula governed by three numbers, the Iwasawa invariants $\lambda$, $\mu$, and $\nu$. These invariants tell you the "law of growth" for the complexity within the entire tower.

And here, the distinction between regular and irregular primes comes roaring back to life. A prime's regularity isn't just a property of the first floor; it has profound implications for the whole skyscraper. For a regular prime $p$, its good behavior is inherited. This is reflected in the Iwasawa invariant $\lambda$ being zero. Let’s take the regular prime $p=5$. When we compute the invariants for the "minus part" of the class group tower, we find that not only is $\lambda=0$ (because 5 is regular) but $\mu$ and $\nu$ are also zero! We see the same for $p=3$. This means that for these regular primes, the $p$-part of the class group does not grow at all as we climb the infinite tower. The simplicity at the bottom guarantees simplicity all the way to the top.

The story gets even deeper. The Iwasawa Main Conjecture, now a celebrated theorem, reveals a breathtaking connection. It states that these algebraic invariants ($\lambda, \mu, \nu$), which describe the growth of class groups, are completely encoded in an analytic object: a $p$-adic L-function. This function is a sophisticated modern cousin of the very Bernoulli numbers that first defined regularity! This is the unity of mathematics in its purest form: an algebraic structure (the tower of class groups) and an analytic object (the $p$-adic L-function) are two sides of the same coin. The algebraic side is "controlled" by elements in the group ring, the so-called Stickelberger elements, which systematically annihilate the class groups, and the properties of these algebraic annihilators are themselves described perfectly by the analytic L-function. A simple test on Bernoulli numbers echoes through algebra and analysis, across an infinite tower of worlds.

You'd be forgiven for thinking that class groups and Iwasawa towers are the exclusive playground of number theorists. But the structures they've uncovered are so fundamental that they appear in the most unexpected places. After all, a class group is a finite abelian group—one of the basic building blocks of modern algebra.

Let's imagine one of these class groups, say for the regular prime $p=29$. We know from number theory that this group is isomorphic to $(\mathbb{Z}/2\mathbb{Z})^3$, a collection of eight elements where every element is its own inverse. Now, let's stop thinking like a number theorist and start thinking like a physicist or a computer scientist. This group is a set of states, a network of nodes. What if we start at the identity element (the principal ideals) and take a "random walk"? At each step, we flip a coin and, based on the outcome, multiply our current position by one of two specific group elements.

We can now ask questions straight out of probability theory: Will we ever get back to where we started? What's the average time to return? The answers depend entirely on the algebraic structure of the group. In this example, one can calculate that you can only return to the identity after an even number of steps, so the "period" of this random walk is 2. The point is not the specific answer, but the connection itself. A problem about the arithmetic of numbers becomes a landscape for modeling a random process. A tool forged to solve Diophantine equations becomes a lattice for a Markov chain.

This is just one example. The ideas and objects that grew out of this lineage—class groups, elliptic curves, modular forms—have found applications and analogues in fields like cryptography, where the difficulty of certain problems in class groups can be used to build secure systems, and error-correcting codes. The quest to understand the humble prime number has, time and again, provided the abstract structures that other branches of science and technology later find indispensable.

So, from a failed, yet brilliant, attack on a 350-year-old problem, the idea of regularity has taken us on an extraordinary journey. It led to the final proof of Fermat’s Last Theorem, a proof that unified vast and disparate areas of mathematics. It evolved into Iwasawa theory, giving us a "calculus" to study infinite towers of number fields and revealing a profound duality between algebra and analysis. And the very objects it studies, like class groups, have become part of the universal toolkit of mathematics, appearing in contexts Kummer could never have imagined.

The legacy of regular primes is not that they solved the problem they were invented for, but that they revealed the problem was far more interesting than anyone thought. They showed us a doorway, and through it, a landscape of breathtaking beauty and interconnectedness. They are a testament to the fact that in mathematics, asking a good question is often more important than finding the immediate answer.