Union of Open Sets in Topology

In the architecture of mathematics, certain rules are so fundamental they act as the load-bearing columns for entire fields. One such rule, found at the heart of topology, governs how we combine the basic building blocks of space known as "open sets." While it may sound simple, the principle that an arbitrary union of open sets is always open is an axiom of immense power and subtlety. This article delves into this cornerstone concept, addressing the question of why this specific property is so crucial. We will first explore the mechanics of this rule in the "Principles and Mechanisms" chapter, contrasting it with intersections and uncovering its elegant duality with closed sets. Following this, the "Applications and Interdisciplinary Connections" chapter will reveal how this single axiom enables mathematicians to sculpt complex spaces, classify infinite structures, and forge connections between topology, analysis, and beyond.

Imagine you are walking in a vast, open field. No matter where you stand, you have a little bit of "wiggle room" in every direction. You can take a tiny step forward, backward, left, or right, and you are still comfortably inside the field. This field is an intuitive picture of what mathematicians call an ​​open set​​. No point inside it is ever "on the edge," because every point comes with its own small, private patch of space surrounding it that is also part of the set. An open interval like $(0, 1)$ has this property. Pick any number inside it, say $0.5$, and you can always find a smaller interval like $(0.49, 0.51)$ around it that is still completely within $(0, 1)$.

Now, what if we had a set like $(0, 1]$, which includes the number $1$? This set is not open. Why? Because if you stand right on the point $1$, you have no wiggle room to the right. Any tiny step to the right, no matter how small, takes you out of the set. The point $1$ is a "hard boundary," a fence post. This simple idea—the existence of wiggle room for every single point—is the seed from which the entire field of topology grows.

Let's return to our fields. Suppose you have a whole collection of these open, fenceless fields. Some might be large, some small, some might even overlap. What happens if you declare all of them to be part of one giant new property? You simply take their ​​union​​. Is this new, larger territory also an open set?

The answer is a resounding yes, and the reasoning is beautifully simple. Pick any point in this grand union. By definition, that point must have come from at least one of the original open fields. And since it was in an open field, it had its own patch of wiggle room. That original field is now entirely contained within your new giant union, so the point's original wiggle room is still there, safe and sound, within the larger set. Nothing has changed for that point. Since this is true for any point you could possibly choose in the union, the union itself must be open.

This isn't just a pleasant intuition; it's a cornerstone of mathematics, a fundamental axiom. It holds true even if you combine an infinite number of open sets. Consider, for example, the set made by taking an infinite collection of open intervals centered at every positive integer: $(1 - \frac{1}{3}, 1 + \frac{1}{3})$, $(2 - \frac{1}{5}, 2 + \frac{1}{5})$, $(3 - \frac{1}{9}, 3 + \frac{1}{9})$, and so on, with the intervals getting smaller and smaller. The full set is $S = \bigcup_{n=1}^\infty ( n - \frac{1}{2^n + 1}, n + \frac{1}{2^n + 1} )$. This collection of disjoint "islands" is still an open set. If you pick a point, it lives in one specific island, say $A_n$, and it inherits its wiggle room from that island. The fact that there are infinitely many other islands doesn't take away its local freedom.

The same principle applies to another interesting set: the real number line with all the integers removed, $\mathbb{R} \setminus \mathbb{Z}$. This set can be seen as the union of all open intervals between consecutive integers: $\dots \cup (-1, 0) \cup (0, 1) \cup (1, 2) \cup \dots$. Since it's a union of open sets, it is itself an open set.

The logic is so fundamental that we can see it at work even in abstract, finite "universes". Imagine a universe with just five points, $X = \{a, b, c, d, e\}$, and we are told that the sets $A_1 = \{a, c\}$, $A_2 = \{b, c, e\}$, and $A_3 = \{d\}$ are open. If we form the union $U = A_1 \cup A_2 \cup A_3 = \{a, b, c, d, e\}$, is it open? To check, we just need to find "wiggle room" for each point.

• For point $a$, its wiggle room is the set $A_1$ it came from.
• For point $b$, its wiggle room is the set $A_2$.
• For point $c$, it's in both $A_1$ and $A_2$, so we can choose either one as its wiggle room.
• For point $d$, its wiggle room is $A_3$.
• For point $e$, its wiggle room is $A_2$. Every single point in the union has a "witness" open set from the original collection that contains it and is fully inside the union. This confirms our principle in a very concrete way: ​​an arbitrary union of open sets is always open​​.

So, unions are straightforward. This might lead you to ask a natural follow-up question: what about intersections? If we take the common area shared by a collection of open sets, is that area also guaranteed to be open?

If we're only dealing with a finite number of sets, the answer is yes. The overlap of two, or three, or a thousand open fields is still an open field. But when we venture into the realm of the infinite, something strange and wonderful happens. The property of openness can vanish.

Let's build a counterexample. Consider an infinite sequence of nested open intervals, each one slightly smaller than the last: $S_1 = (-1, 2)$ $S_2 = (-\frac{1}{2}, 1 + \frac{1}{2})$ $S_3 = (-\frac{1}{3}, 1 + \frac{1}{3})$ ... $S_n = (-\frac{1}{n}, 1 + \frac{1}{n})$ ... Each of these sets, $S_n$, is perfectly open. Now, what is their intersection? What is the set of points that belong to every single one of these intervals?

Let's think about the boundaries. The left endpoint, $-\frac{1}{n}$, gets closer and closer to $0$ from the left. The right endpoint, $1 + \frac{1}{n}$, gets closer and closer to $1$ from the right. A point like $-0.001$ will eventually be kicked out when $n$ is large enough (e.g., for $n=1001$, $-\frac{1}{1001} \gt -0.001$). A point like $1.001$ will also be kicked out eventually. The only points that manage to stay in every single set are the ones in the interval $[0, 1]$. The result of this infinite intersection is $\bigcap_{n=1}^{\infty} S_n = [0, 1]$.

And the set $[0, 1]$ is a ​​closed interval​​. It is not open! The points $0$ and $1$ are hard boundaries; they have no wiggle room. We started with an infinite collection of open sets, and their intersection collapsed into something that isn't open at all. This reveals a profound asymmetry in the rules of topology: openness is preserved under arbitrary unions, but only under finite intersections.

This asymmetry might seem like an odd quirk, but it's actually a clue to a deeper, more beautiful structure. To see it, we must introduce the concept of a ​​closed set​​. A set is defined as closed if its complement—everything not in the set—is open. The closed interval $[0,1]$ is a closed set because its complement, $(-\infty, 0) \cup (1, \infty)$, is a union of two open intervals and is therefore open.

This simple definition, connecting closed and open sets through complements, is like a magic mirror. Every property of open sets has a corresponding "mirror image" property for closed sets, thanks to a powerful logical tool called ​​De Morgan's Laws​​. These laws tell us that the complement of a union is the intersection of the complements, and the complement of an intersection is the union of the complements.

Let's apply this to what we know. We know that an ​​arbitrary intersection of closed sets is always closed​​. How can we be so sure? Let's take any collection of closed sets, $\{F_i\}$. Their intersection is $\bigcap F_i$. By De Morgan's Law, the complement of this intersection is the union of the complements: $\left( \bigcap_i F_i \right)^c = \bigcup_i F_i^c$ Since each $F_i$ is closed, its complement $F_i^c$ is open. The expression on the right is an arbitrary union of open sets, which we already established is always open. So, $(\bigcap_i F_i)^c$ is open. And if the complement of a set is open, the set itself must be closed! The logic is inescapable.

What about unions of closed sets? The mirror logic applies. We saw that infinite intersections of open sets can fail to be open. The mirror image of this is that infinite unions of closed sets can fail to be closed. For example, the union of the closed singleton sets $\{1\}$, $\{\frac{1}{2}\}$, $\{\frac{1}{3}\}, \dots$ is the set $\{1, \frac{1}{2}, \frac{1}{3}, \dots \}$. This set is not closed because the point $0$ is a limit point of the set (you can get arbitrarily close to it), but $0$ is not in the set itself.

This creates a beautiful duality:

Open Sets	Closed Sets
​​Arbitrary union​​ is open.	​​Finite union​​ is closed.
​​Finite intersection​​ is open.	​​Arbitrary intersection​​ is closed.

The axioms of topology can be expressed entirely in the language of open sets or, with equal validity, in the language of closed sets. They are two sides of the same coin, elegantly connected by the concept of the complement.

With these fundamental principles in hand, we can construct more sophisticated ideas. A very useful concept is the ​​interior​​ of a set $A$, written as $\text{int}(A)$. The interior is the largest possible open set that you can fit entirely inside $A$. It's what's left of $A$ after you've shaved off all its boundary points.

How would one construct such a set? The most direct way is to gather up all the open sets that are contained in $A$ and take their union. $\text{int}(A) = \bigcup \{ O \mid O \text{ is open and } O \subseteq A \}$ Now, we can ask a crucial question: is the interior of a set, $\text{int}(A)$, itself always an open set? Based on our very first principle, the answer is immediate and obvious. Since $\text{int}(A)$ is defined as a union of open sets, it must be open! This is a perfect example of how a simple, powerful axiom provides effortless answers to seemingly complex questions. The structure of mathematics propagates from these simple rules, revealing a coherent and interconnected world. The notion that an arbitrary union of open sets is open is not just one rule among many; it is a foundational mechanism that breathes life into the very definition of space.

We have just learned a rule that seems almost trivially simple: if you take any collection of open sets—two, a thousand, or an infinite collection as vast as the real numbers themselves—and you combine them through union, the resulting set is still, unfailingly, open. Is this just a dry, axiomatic statement for mathematicians to file away? Far from it. This single rule is a golden thread that weaves through the fabric of modern mathematics. It is a tool of immense power and flexibility, allowing us to build, dissect, and understand the very nature of space. Let's go on an adventure to see where this simple idea takes us, from the familiar landscapes of our world to the abstract frontiers of mathematical thought.

Let's begin in the familiar territory of two-dimensional space, the plane of a graph or a map. Imagine a smooth, curving line, perhaps the path of a particle described by the parabola $y = x^2$. What if we wanted to describe not just the path itself, but a "safety corridor" around it? We could do this by placing a small open disk, like a tiny circular force field, around every single point on the parabola. The resulting region would be the union of all these disks—an uncountable infinity of them! Our axiom provides a wonderful guarantee: because each disk is an open set, their grand union is also an open set. Any point inside this corridor has some "wiggle room" before it hits the boundary. This method of "thickening" a set is incredibly powerful, allowing us to define neighborhoods and regions of influence in a rigorous way.

We can also build more complex open sets by piecing together simpler ones. Consider an infinite staircase made of open squares, each one occupying a space like $(n, n+1) \times (n, n+1)$ for every integer $n$. The union of all these disjoint squares forms a single, large, but rather strangely shaped open set. The axiom holds firm: a union of open sets, even a countably infinite one, is open. This has an immediate and important consequence for the set's complement—the grid of lines and vertices that forms the boundary of these squares. Because the union of squares is open, its complement must be a closed set. This beautiful duality between open and closed, between union and intersection, is a cornerstone of topology, and it all hinges on our simple rule.

This principle extends beyond geometric shapes. Consider a function like $f(x) = \cos(1/x)$. The points where the function's value is, say, less than $1/2$ form a rather complicated set of intervals near zero. Yet, we can be certain this set is open. Why? Because the function is continuous (away from zero), and we are essentially looking at the preimage of the open interval $(-\infty, 1/2)$. Continuity ensures that points close to a solution are also solutions, which is the very essence of an open set. In other cases, an open set might be presented to us directly as a union of an infinite number of open intervals, and we immediately know it is open without any further checks. The axiom does the work for us.

So far, we have used the rule to analyze sets within a pre-existing space like $\mathbb{R}^n$. But its power runs deeper. The axiom is a fundamental tool for constructing topological spaces from scratch.

Imagine a set $X$ with no structure at all, just a collection of points. What if we make a radical declaration: every single point, enclosed in its own set $\{x\}$, is an open set. What happens now? Our axiom of unions springs into action. If we want to form any subset $A \subseteq X$, no matter how complicated, we can do so by simply taking the union of all the singleton open sets $\{x\}$ for every point $x$ in $A$. Since this is a union of open sets, the resulting set $A$ must also be open. In this "discrete topology," every single subset is open! This demonstrates the immense constructive power of allowing arbitrary unions. With the simplest possible open components (single points), we can build the most complex topology imaginable—one where every subset is open—all thanks to the union axiom. This shows the rule is not just descriptive; it is generative.

The implications of this axiom ripple out far beyond topology itself, shaping entire fields like measure theory and analysis. In measure theory, one seeks to define a consistent notion of "size" or "length" or "volume" for as many sets as possible. A natural first thought might be to work with the collection of all open sets. They seem well-behaved. Indeed, this collection is closed under countable unions, as guaranteed by our axiom.

However, a strange thing happens when we consider complements. The complement of an open interval like $(0, 1)$ is the set $(-\infty, 0] \cup [1, \infty)$, which is not open because it includes its boundary points. So, the collection of all open sets is not closed under complementation. This "failure" is profoundly important. It tells us that the collection of open sets, while beautifully structured with respect to unions, is not robust enough for the purposes of measure theory, which requires closure under complements and countable unions (a so-called $\sigma$-algebra). This very observation motivates the construction of a richer collection, the Borel sets, which is the smallest $\sigma$-algebra containing all the open sets. The journey to the powerful theory of Lebesgue measure begins with appreciating what the collection of open sets can do (unions) and what it cannot (complements).

The union principle also gives us a crucial tool for dissection. For any set $E$, its interior, denoted $E^\circ$, is defined as the largest open set contained within it. How is such a set found? Simple: it is the union of all open sets contained in $E$. Our axiom guarantees that this union is itself an open set. This is a gift. It means that no matter how pathological or strange a set $E$ might be, its interior is always a nice, well-behaved open set. This has direct consequences for measurability. Since all open sets are Lebesgue measurable, the interior of any set is always Lebesgue measurable. This is not a deep theorem to be proven with difficulty; it is an immediate and elegant consequence of how we define the interior using the union of open sets.

The axiom of unions is also central to how we classify and analyze the structure of infinite spaces. Many important spaces, like the real line $\mathbb{R}$, are not compact, which can make them difficult to work with. However, we can often tame this unwieldy infinity by expressing the space as a union of simpler, more manageable pieces. For instance, the real line can be written as the countable union of open intervals $\mathbb{R} = \bigcup_{n=1}^\infty (-n, n)$. Each interval $(-n, n)$ is not compact, but its closure $[-n, n]$ is. Spaces that can be written as a countable union of open sets with compact closures are fundamental in analysis. This property, known as being a countable union of relatively compact open sets (which implies $\sigma$-compactness in a locally compact Hausdorff space), allows us to extend results proven on "nice" compact sets to the entire, non-compact space. The union principle is the engine that drives this powerful technique of approximation from within.

We can even classify spaces based on how they behave with respect to unions. A space is called Lindelöf if any open cover of the whole space has a countable subcover. But an even stronger property, called hereditarily Lindelöf, can be characterized in a breathtakingly simple way using our axiom. A space is hereditarily Lindelöf if and only if for any collection of open sets, their union can be formed by a countable sub-collection. This means that in such spaces (which include all of $\mathbb{R}^n$), the complexity of forming unions is fundamentally tamed: any union, even one over an uncountably infinite index set, is equivalent to a much simpler countable union. This deep structural fact, which governs the very texture of the space, is phrased entirely in the language of our axiom.

Finally, let's see how our axiom guides us in constructing topologies on more exotic, abstract structures, like those found in algebraic topology. Imagine an infinite line built from discrete vertices and the edges connecting them. We want to put a topology on this object. What does it mean for a set to be open here? One clever way is to define the "weak topology," where a set is declared open if its intersection with every finite piece of the structure is open. Now, consider the set of all "open edges"—the edges without their endpoints. Is this set open in our infinite line? Yes. Its intersection with any finite collection of edges is just a finite union of open intervals, which is open. Therefore, by definition, the infinite union is open in the whole space. We have engineered a topology on an infinite object specifically so that our intuition about unions continues to hold, allowing us to build a consistent and workable theory.

So, the next time you think of an "open set," don't just picture an empty room or an interval on a line. Picture a dynamic, flexible entity, defined by a principle of boundless combination. This one axiom is not a restriction, but a license to create. It is the glue that holds together the shapes of our mathematical imagination, revealing a profound and beautiful unity across the vast landscape of mathematics.