Completions of Fields

The concept of completion is a fundamental tool in mathematics, most familiarly seen in the construction of the real numbers from the rationals to 'fill the gaps' like $\sqrt{2}$. But this process hinges entirely on our standard notion of distance. What if there were other, equally valid ways to measure the gap between numbers? This question reveals a vast landscape of new number systems, each with its own unique properties and geometry. This article explores the profound idea of field completion. In the first part, "Principles and Mechanisms", we will construct these new worlds, defining the p-adic numbers and exploring their bizarre ultrametric geometry. In the second part, "Applications and Interdisciplinary Connections", we will demonstrate how these constructed fields are not mere curiosities but powerful tools for solving long-standing problems in number theory, from finding roots of polynomials to understanding the very nature of rational solutions to equations.

Imagine the rational numbers, $\mathbb{Q}$, the familiar world of fractions. It seems complete, ordered, and well-behaved. Yet, as the ancient Greeks discovered to their dismay, it is full of "holes." Numbers like $\sqrt{2}$ or $\pi$, which can be approximated ever more closely by fractions, are not themselves fractions. To fill these gaps, we "complete" the rational numbers, creating the continuous, seamless line of real numbers, $\mathbb{R}$. This process of completion is like turning a dotted line into a solid one, by adding the limit points of every converging sequence.

But what if our very notion of "distance" and "closeness" was different? What if there were other, equally valid ways to measure the size of a number? This is the central idea that unlocks a vast and beautiful landscape of new number systems. The principles of completion are not just about one process, but about a universal tool that allows mathematicians to build new worlds from old ones, each revealing a different facet of the underlying structure.

Our everyday understanding of a number's "size" is its distance from zero on the number line—its absolute value. For a rational number $x$, we denote this by $|x|_{\infty}$. This familiar distance measure obeys the triangle inequality: $|x+y|_{\infty} \le |x|_{\infty} + |y|_{\infty}$. But is this the only way?

Let's pick a prime number, say $p=5$. Instead of asking how "big" a number is, let's ask how "divisible by 5" it is. A number like $50 = 2 \times 5^2$ is more divisible by 5 than $15 = 3 \times 5^1$, which is in turn more divisible than $6$. We can capture this with the ​​$p$-adic valuation​​, denoted $v_p(x)$, which is simply the exponent of $p$ in the prime factorization of $x$.

For example, with $p=5$:

• $v_5(50) = 2$
• $v_5(15) = 1$
• $v_5(6) = 0$
• $v_5(\frac{3}{25}) = v_5(3 \times 5^{-2}) = -2$

Now, let's define a new "size," the ​​$p$-adic absolute value​​, which we'll denote $|x|_p$. We want numbers that are highly divisible by $p$ to be considered "small." A natural way to do this is to define $|x|_p = p^{-v_p(x)}$, with $|0|_p=0$ by convention. Let's see what this does for our examples:

• $|50|_5 = 5^{-2} = \frac{1}{25}$ (very small)
• $|15|_5 = 5^{-1} = \frac{1}{5}$ (small)
• $|6|_5 = 5^{-0} = 1$ (not small)
• $|\frac{3}{25}|_5 = 5^{-(-2)} = 25$ (very large!)

In this strange new world, a number is small if it contains many factors of $p$. Two numbers are "close" if their difference is divisible by a high power of $p$.

This new way of measuring distance has bizarre and wonderful consequences. It doesn't just satisfy the familiar triangle inequality, but a much stronger one called the ​​ultrametric inequality​​: $|x+y|_p \le \max\{|x|_p, |y|_p\}$ This seems like a small change, but it warps geometry into something completely unrecognizable.

Imagine a triangle with side lengths $a=|x|_p$, $b=|y|_p$, and $c=|x+y|_p$. The ultrametric inequality tells us that the length of any side is no greater than the longer of the other two. This leads to a startling conclusion: in any triangle, the two longest sides must be of equal length! So, in a $p$-adic world, all triangles are isosceles.

The geometric features are just as strange. In the world of real numbers, a circle has a unique center. In a $p$-adic world, any point inside a ball is its center. Furthermore, these balls are both open and closed sets at the same time ("clopen"). This means there are no continuous paths from inside a ball to outside it without a "jump." The entire space is ​​totally disconnected​​, like a fine dust of points, in stark contrast to the connected, continuous real line $\mathbb{R}$.

Just as the rational numbers $\mathbb{Q}$ have gaps from the perspective of the usual distance, they also have gaps when viewed through the lens of a $p$-adic distance. The process of completion is the same: we take all sequences of rational numbers that are "Cauchy"—meaning the terms get arbitrarily close to each other—and for each sequence that doesn't already converge to a rational number, we "invent" a new number to be its limit.

When we do this for the usual absolute value $| \cdot |_{\infty}$, we get the real numbers $\mathbb{R}$.

When we do this for the $p$-adic absolute value $| \cdot |_p$, we create a completely new field: the field of ​​$p$-adic numbers​​, denoted $\mathbb{Q}_p$. And we can do this for every prime $p$, giving us an infinite family of new, complete worlds: $\mathbb{Q}_2, \mathbb{Q}_3, \mathbb{Q}_5$, and so on.

What do these $p$-adic numbers look like? They can be thought of as power series in $p$. For instance, in the field of $5$-adic numbers $\mathbb{Q}_5$, consider the equation $x^2 + 1 = 0$. This has no solution in the real numbers. But in $\mathbb{Q}_5$, we can find a solution! Let's try to build one.

• Modulo 5: we need $x^2 \equiv -1 \equiv 4 \pmod 5$. We can choose $x_0 = 2$.
• Modulo 25: we look for a solution of the form $x_1 = 2 + 5k$. The condition $(2+5k)^2+1 \equiv 0 \pmod{25}$ simplifies to $5+20k \equiv 0 \pmod{25}$, which gives $1+4k \equiv 0 \pmod 5$, so $k=1$. Our next approximation is $x_1 = 7$.
• Modulo 125: we look for $x_2 = 7 + 25j$. The condition $(7+25j)^2+1 \equiv 0 \pmod{125}$ simplifies to $50+350j \equiv 0 \pmod{125}$, which gives $2+14j \equiv 0 \pmod 5$, so $j=2$. Our next approximation is $x_2 = 57$.

This process can be continued indefinitely, producing a sequence $2, 7, 57, \ldots$. In the $5$-adic sense, this sequence is converging because the difference between successive terms is divisible by ever-higher powers of 5. For example, $|57-7|_5 = |50|_5 = |2 \times 5^2|_5 = \frac{1}{25}$. The limit of this sequence is a true $5$-adic number $\xi = 2 + 1 \cdot 5 + 2 \cdot 5^2 + \dots$, and in the world of $\mathbb{Q}_5$, it is a genuine square root of $-1$.

We have seen two fundamentally different ways to complete the rational numbers $\mathbb{Q}$: one leading to the familiar real numbers $\mathbb{R}$, and another leading to an infinite family of exotic $p$-adic fields $\mathbb{Q}_p$. A natural question arises: are there any other ways? Could we invent a "7.5-adic" absolute value, or something even stranger?

The answer is a breathtakingly simple and profound "No." A celebrated result known as ​​Ostrowski's Theorem​​ states that every non-trivial absolute value on the field of rational numbers $\mathbb{Q}$ is equivalent to either the usual absolute value $|\cdot|_{\infty}$ or a $p$-adic absolute value $|\cdot|_p$ for some prime $p$.

This theorem is a cornerstone of modern number theory. It tells us that our two types of completions, $\mathbb{R}$ and the various $\mathbb{Q}_p$, are not just a random collection of constructions. They are, in a very precise sense, the only ways to complete the rational numbers. Each completion provides a different "lens" through which to view the arithmetic of fractions. The real numbers care about size and order. The $p$-adic numbers care about divisibility by $p$. To get a full picture of a problem in number theory—a "global" understanding—one often needs to look at it "locally" through all of these lenses at once.

This powerful idea of completion is not limited to the rational numbers. It is a universal principle that applies to a wide class of fields. Consider, for instance, the field of rational functions $\mathbb{Q}(x)$, which consists of ratios of polynomials with rational coefficients.

We can define a valuation on this field analogous to the $p$-adic valuation. Instead of a prime number, let's use the variable $x$. The valuation $v_0(f)$ measures the order of the zero or pole of the function $f(x)$ at $x=0$. For example, $v_0(x^3) = 3$, $v_0(1/x^2) = -2$, and $v_0(\frac{x+1}{x-1}) = 0$.

Completing the field $\mathbb{Q}(x)$ with respect to this valuation gives us the field of ​​formal Laurent series​​, $\mathbb{Q}((x))$. An element of this field is a series of the form $\sum_{k=n}^{\infty} a_k x^k$, where the series can have finitely many negative powers of $x$ but extends infinitely in the positive direction. This is the natural home for the Taylor and Laurent series you encounter in calculus! A sequence of polynomials that agree on more and more terms is a Cauchy sequence, and its limit is a formal power series. This construction mirrors the creation of $p$-adic numbers, where an integer is represented as a series in powers of $p$.

This reveals a deep unity. The general objects of study are ​​global fields​​, which fall into two main classes: number fields (finite extensions of $\mathbb{Q}$) and global function fields (like the field of functions on a curve over a finite field). For any such field $K$, there is a set of distinct ways to measure size, called ​​places​​. Each place $v$ corresponds to an absolute value $|\cdot|_v$ and gives rise to a completion $K_v$, which is a ​​local field​​.

Just as with $\mathbb{Q}$, these local fields fall into two categories.

1. ​​Archimedean completions​​: These arise from places analogous to the usual absolute value. The resulting fields are always isomorphic to the real numbers $\mathbb{R}$ or the complex numbers $\mathbb{C}$.
2. ​​Non-Archimedean completions​​: These arise from places analogous to the $p$-adic valuations. The resulting fields have ultrametric geometry, like the $p$-adic fields or fields of formal Laurent series.

This stunning framework unifies the study of numbers and functions. Completing a a field is like focusing a microscope on a single point, revealing the intricate "local" structure that is invisible from a "global" perspective. By piecing together the information from all these local pictures—from all the different completions—we can solve problems and uncover structures that would otherwise remain hidden. It is a testament to the power of changing one's perspective, and to the beautiful, unified fabric of mathematics.

Having journeyed through the formal construction of completed fields, one might reasonably ask: what is all this for? Are these $p$-adic numbers and their kin merely a mathematician's fanciful creation, an esoteric parallel universe with no bearing on the "real" world of numbers we know? The answer, you will be delighted to find, is a resounding no. The process of completion is not an act of escapism but a profound analytical tool. It allows us to put a problem under a new kind of microscope, to view it "locally" at each prime $p$, and by assembling these local pictures, to gain an incredibly sharp image of the global whole. In this chapter, we explore the astonishingly diverse applications of this idea, from solving ancient number theory problems to describing the geometry of fractals.

Let's begin with a puzzle. What is the sum of the series $1 \cdot 1! + 2 \cdot 2! + 3 \cdot 3! + \dots$? In the world of real numbers, the terms $n \cdot n!$ explode towards infinity, and the series diverges faster than almost anything you can imagine. It is utterly meaningless. But what if I told you that in the world of 7-adic numbers, this series not only converges, but its sum is exactly $-1$? That is, $\sum_{n=1}^\infty n \cdot n! = -1 \quad (\text{in } \mathbb{Q}_7)$ This result, which can be verified with a beautiful "telescoping sum" argument, is no mere parlor trick. It's a direct consequence of the bizarre, "ultrametric" geometry of the $p$-adic world. Here, a series converges simply if its terms go to zero. For the 7-adic valuation, the term $(n+1)!$ contains many more powers of 7 than $n!$, so its "size" $|(n+1)!|_7$ shrinks rapidly. The arithmetic in these fields, while following the same basic rules of addition and multiplication, operates under a completely different notion of "closeness," leading to results that are at once deeply foreign and rigorously correct.

So, how does one actually compute in this strange new world? How, for instance, could we find a number whose square is 11 in the field of 5-adic numbers, $\mathbb{Q}_5$? It turns out the process is remarkably algorithmic and intuitive, bearing a striking resemblance to the numerical methods you might have learned in calculus. We can build the solution, a 5-adic integer $x = d_0 + d_1 \cdot 5 + d_2 \cdot 5^2 + \dots$, digit by digit.

First, we look for a solution just "modulo 5": we need $d_0^2 \equiv 11 \equiv 1 \pmod{5}$. This gives us two choices, $d_0=1$ or $d_0=4$. Let's pick $d_0=1$. Next, we refine this guess. We look for the next digit, $d_1$, such that $(1 + d_1 \cdot 5)^2 \equiv 11 \pmod{25}$. A little algebra shows that $d_1$ must be $1$. We can continue this process indefinitely, at each stage lifting our solution from modulo $5^k$ to a more precise solution modulo $5^{k+1}$. This beautiful step-by-step refinement is the essence of a profoundly powerful tool called ​​Hensel's Lemma​​, which is the $p$-adic analogue of Newton's method for finding roots. It tells us that if we can find an approximate solution modulo $p$, we can often lift it uniquely to an exact solution in the full $p$-adic field.

This power to solve equations extends beyond just finding numbers; it helps us understand the very structure of new fields we might build. For example, if we ask whether we can adjoin $\sqrt{3}$ to the 7-adic numbers, we are asking if the polynomial $x^2-3=0$ has a solution in $\mathbb{Q}_7$. Hensel's Lemma tells us to first check modulo 7. A quick check reveals that 3 is not a perfect square modulo 7. There is no integer whose square is $3, 10, 17, \dots$. This simple fact from grade-school arithmetic is enough to prove that $\sqrt{3}$ does not exist in $\mathbb{Q}_7$, and therefore adjoining it creates a genuine field extension of degree 2. The local, completed world is built upon the bedrock of finite, modular arithmetic.

Hensel's Lemma is great for finding roots, but what if we want to know more? What if we want to know the size—the $p$-adic valuation—of the roots of a polynomial, even without finding them explicitly? For this, mathematicians have devised a wonderfully elegant visual tool: the ​​Newton Polygon​​.

Suppose we have a polynomial like $P(x) = x^4 + px^3 + p^3x + p^7$. To understand its roots in the algebraic closure of $\mathbb{Q}_p$, we translate its structure into a simple picture. On a graph, we plot points $(i, v_p(a_i))$ for each term $a_i x^i$—the x-coordinate is the power of the variable, and the y-coordinate is the p-adic valuation of the coefficient. For our example, we would plot $(4, 0)$, $(3, 1)$, $(1, 3)$, and $(0, 7)$. Now, imagine tying a string to the lowest point on the y-axis and wrapping it around the points from below. The resulting shape is the Newton polygon.

The magic is this: the slopes of the segments of this polygon tell you the valuations of the roots! A segment of a certain slope, say $-m$, with a horizontal length of $L$, corresponds to exactly $L$ roots having a $p$-adic valuation of $m$. For our polynomial $P(x)$, the polygon has a segment of slope $-4$ and length 1, and another of slope $-1$ and length 3. This tells us, with no further calculation, that the polynomial has one root with valuation 4 and three roots with valuation 1. The Newton polygon is a beautiful example of how a geometric perspective can reveal deep algebraic truths, like an X-ray for polynomials over local fields.

The true power of completions comes not from studying any single $\mathbb{Q}_p$ in isolation, but by considering all of them—including the real numbers $\mathbb{R}$, the completion at the "infinite prime"—all at once. This philosophy is enshrined in the ​​Local-Global (or Hasse) Principle​​. The principle poses a profound question: if an equation has a solution in every completion of $\mathbb{Q}$ (i.e., in $\mathbb{R}$ and in every $\mathbb{Q}_p$), must it have a solution in $\mathbb{Q}$ itself?

Think of it like a jigsaw puzzle. A global solution over $\mathbb{Q}$ is the complete picture. The existence of a solution in each completion $K_v$ is like knowing that each individual puzzle piece exists. Does the existence of all the pieces guarantee they can be fit together?

For some problems, the answer is a glorious "yes!" The most famous example is the ​​Hasse-Minkowski Theorem​​, which states that a quadratic equation, like $a_1x_1^2 + \dots + a_nx_n^2 = 0$, has a rational solution if and only if it has a solution in $\mathbb{R}$ and in every $\mathbb{Q}_p$. This is an incredibly powerful result. It reduces a single, hard question about the infinite field of rational numbers to a series of simpler, often algorithmic, questions about local fields.

However, the world of numbers is subtle. The Hasse Principle is not a universal law. Famously, it fails for certain cubic equations. The equation $3x^3 + 4y^3 + 5z^3 = 0$, for example, has solutions in $\mathbb{R}$ and in every field $\mathbb{Q}_p$, yet it has no non-trivial solution in the rational numbers $\mathbb{Q}$. This "failure" is not a defeat; it is the gateway to deeper mathematics. The obstruction that prevents the local solutions from being pieced together into a global one is measured by a sophisticated object called the Tate-Shafarevich group, a central player in the modern theory of elliptic curves.

The local-global principle begs for a language that can speak about all completions simultaneously. That language is provided by the ​​ring of adeles​​, $\mathbb{A}_{\mathbb{Q}}$. This remarkable structure is a "restricted product" of all the completions of $\mathbb{Q}$. An adele is an infinite vector $(x_\infty, x_2, x_3, x_5, \dots)$, with a component in each completion $\mathbb{Q}_v$. The crucial "restriction" is that for all but a finite number of primes $p$, the component $x_p$ must be a $p$-adic integer. This condition keeps the object from becoming pathologically large, making it a locally compact ring that contains the rational numbers as a discrete subring.

The adele ring is the ultimate expression of the local-global philosophy. It provides a single, unified stage on which the real numbers and all the $p$-adic numbers interact. The algebraic justification for this beautiful structure lies in the fact that for any number field $K$, tensoring with $\mathbb{Q}_p$ breaks the field apart into its local constituents above $p$: $K\otimes_{\mathbb{Q}}\mathbb{Q}_{p} \cong \prod_{\mathfrak{p}|p} K_{\mathfrak{p}}$. The adele ring essentially performs this decomposition at all primes at once. It is the natural setting for modern algebraic number theory, from the study of reciprocity laws to the proof of Fermat's Last Theorem.

The idea of completion is not confined to the rational numbers, nor are its applications limited to number theory. The same process can be applied to any number field, such as the Gaussian rationals $\mathbb{Q}(i)$, by completing at the prime ideals of its ring of integers. This provides a vast and consistent framework for studying arithmetic across all of algebraic number theory.

Furthermore, these ideas forge surprising connections to other branches of mathematics and science. Consider a question from the intersection of number theory and fractal geometry: what is the "size" of the set of real numbers $x$ that can be simultaneously well-approximated by rational numbers in both the real metric and, say, the 5-adic metric? That is, what is the dimension of the set of $x \in [0,1]$ for which there are infinitely many fractions $p/q$ satisfying both $|x - p/q| q^{-3}$ and $|x - p/q|_5 q^{-1}$? The answer, it turns out, is a fractal dimension of $1/2$. This field, known as metric Diophantine approximation, uses the interplay between different completions to define and measure intricate, fractal-like subsets of the real line.

The journey of completions takes us from a surprising calculation to the deepest structural theorems of number theory, and out into the frontiers of modern mathematics. Far from being an idle abstraction, the completion of fields is a lens of unparalleled power, revealing hidden structures and unifying disparate worlds in a single, beautiful mathematical framework.