Pasting Lemma

In mathematics, how do we build complex, continuous structures from simple, well-understood pieces? When piecing together different functions, a fundamental challenge arises: ensuring the final construction is seamless, without any abrupt jumps or tears at the boundaries. The ​​Pasting Lemma​​ provides an elegant and powerful answer to this very problem, serving as a master rule for "stitching" functions together while preserving the crucial property of continuity. It is one of the most practical tools in the study of topological spaces, turning the intuitive act of gluing into a rigorous mathematical certainty.

This article explores the Pasting Lemma from its core principles to its profound applications. First, in the "Principles and Mechanisms" chapter, we will unpack the mechanics of the lemma, exploring the intuition behind it, the critical role of closed sets in its formal statement, and how it applies to dynamic processes like deformations over time. We will also distinguish it from related concepts to clarify its specific purpose. Following this, the "Applications and Interdisciplinary Connections" chapter will demonstrate how this seemingly simple tool becomes the architectural cornerstone for major concepts in algebraic topology, enabling the construction of paths, homotopies, and even the algebraic structure of the fundamental group.

Imagine you are a master tailor, but instead of fabric, you work with the very essence of mathematical functions. Your goal is to stitch together different pieces of continuous "cloth" to create a single, seamless garment. How do you ensure that the final creation has no rips, no sudden jumps, no unsightly puckers at the seams? This is the fundamental question that the ​​Pasting Lemma​​ answers, and it is one of the most practical and elegant tools in the study of continuous spaces.

Let's start with the simplest possible scenario. You have the real number line, $\mathbb{R}$, and you cut it in two at a single point, say at $x=1$. You now have two closed pieces: the ray $(-\infty, 1]$ and the ray $[1, \infty)$. On each piece, you define a perfectly smooth, continuous function. For example, on the left piece you might have $g(x) = \arctan(x)$, and on the right, $h(x) = \frac{\pi}{4} x^2$. How do you create a single continuous function $f(x)$ for the entire line?

The answer is beautifully simple: the two pieces of "cloth" must meet perfectly at the seam. The value of the left function at the seam point, $g(1)$, must be identical to the value of the right function at the seam point, $h(1)$. In our example, $\arctan(1) = \frac{\pi}{4}$ and $\frac{\pi}{4}(1)^2 = \frac{\pi}{4}$. They match! Because the values agree at the point of intersection, the resulting function is continuous across the entire real line. If they didn't match, say if the right-hand function were $3x-1$, then at $x=1$ the left side would still be $\arctan(1) = \frac{\pi}{4}$ while the right would be $3(1)-1=2$. At that point, the function would be torn apart—a discontinuity.

This idea scales up in fascinating ways. What if our seam isn't just a single point, but a whole curve? Imagine the plane $\mathbb{R}^2$. Let's take two closed regions: the unit disk $A = \{(x,y) \mid x^2 + y^2 \le 1\}$ and everything outside it, $B = \{(x,y) \mid x^2 + y^2 \ge 1\}$. The "seam" where they meet is their intersection, the unit circle $S^1 = \{(x,y) \mid x^2+y^2=1\}$.

Now, suppose we have a continuous function $g$ defined on the disk and another continuous function $h$ defined on the exterior. To paste them into one continuous function on the whole plane, the condition is the same, but much more demanding: $g(p)$ must equal $h(p)$ for every single point $p$ on the unit circle. For instance, if on the disk we have $g(x,y) = x(x^2+y^2)$ and on the exterior we have $h(x,y)=x$, then on the seam where $x^2+y^2=1$, the first function becomes $g(x,y) = x(1)=x$, which is identical to the second function $h(x,y)=x$. The seam is perfect, and the resulting global function is continuous.

However, if the functions were, say, $g(x,y) = x^2 - y^2$ and $h(x,y) = x - y$, a quick check at the point $(\frac{1}{\sqrt{2}}, -\frac{1}{\sqrt{2}})$ on the circle reveals a problem. The first function gives $0$, while the second gives $\sqrt{2}$. The seam doesn't hold. The values must match everywhere along the intersection for the gluing to be successful. This principle is so strict that we can use it to solve for unknown parameters in our functions to force them to be continuous.

What we have been exploring is the intuition behind the ​​Pasting Lemma​​. Formally, it states:

Let a topological space $X$ be the union of two ​​closed​​ subsets, $A$ and $B$. If $f: A \to Y$ and $g: B \to Y$ are continuous functions that agree on the intersection $A \cap B$ (that is, $f(p) = g(p)$ for all $p \in A \cap B$), then the function $h: X \to Y$ defined by pasting them together is continuous.

The word ​​closed​​ is doing a lot of work here. Why is it so important? A function is continuous if the inverse image of any closed set in the target space is a closed set in the domain. When we try to prove our pasted function $h$ is continuous, we take a closed set $C$ in $Y$ and look at its inverse image, $h^{-1}(C)$. This inverse image is the union of the inverse images from each piece: $f^{-1}(C) \cup g^{-1}(C)$. Because $f$ and $g$ are continuous on their respective domains, $f^{-1}(C)$ is closed in $A$, and $g^{-1}(C)$ is closed in $B$. And since $A$ and $B$ are themselves closed in the larger space $X$, it follows that these pieces are also closed in $X$. The union of two closed sets is always closed, so $h^{-1}(C)$ is closed, and our pasted function $h$ is continuous. The property of "closedness" is preserved through the operation.

Interestingly, an analogous lemma holds if the two subsets $A$ and $B$ are both ​​open​​. The logic is nearly identical, but it relies on the property that the inverse image of an open set is open. This shows a wonderful symmetry in the topological world. In fact, this pasting principle is quite robust; it even works for other properties, like constructing a ​​closed map​​ (a function that sends closed sets to closed sets) from two smaller closed maps.

The Pasting Lemma isn't just for static functions; it can be applied to dynamic processes. In topology, we study the idea of ​​homotopy​​, which is a formal way of talking about continuously deforming one function into another. Think of a homotopy as a movie: it's a continuous function $H(p, t)$ that depends on a point in space, $p$, and a time parameter, $t$, which runs from $0$ to $1$. At time $t=0$, you have your starting function, $H(p,0) = f(p)$, and at time $t=1$, you have your final function, $H(p,1) = g(p)$.

Can we paste homotopies? Imagine trying to deform a function on the northern hemisphere of a sphere while simultaneously deforming another function on the southern hemisphere. To get a single, continuous deformation of the whole sphere, the Pasting Lemma must apply at every instant in time. The deformation happening along the equator (the seam) must be identical for both the northern and southern "movies" at all times $t \in [0,1]$.

For example, suppose we want to see if the constant map $f(p)=(1,0)$ on the sphere $S^2$ is homotopic to the constant map $g(p)=(-1,0)$. One might try to construct a homotopy on the northern hemisphere that rotates the point along the top half of a circle, $F(p,t) = (\cos(\pi t), \sin(\pi t))$, and another on the southern hemisphere that rotates it along the bottom half, $G(p,t) = (\cos(\pi t), -\sin(\pi t))$. Both are valid homotopies on their own domains. But when we check the seam (the equator), they don't agree for any time between $t=0$ and $t=1$. Trying to paste them would "tear" the sphere apart during the deformation. This specific attempt at pasting fails.

However, this failure doesn't mean the task is impossible! It just means our specific method of gluing was flawed. A much simpler global homotopy exists: just let the point move along the circle for all points on the sphere simultaneously, $H(p,t) = (\cos(\pi t), \sin(\pi t))$. This works perfectly. This teaches us a profound lesson: the Pasting Lemma gives us a method for constructing continuous functions, but if the construction fails, it doesn't rule out the existence of the object we seek.

It's crucial to understand what a tool is for and, just as importantly, what it is not for. The Pasting Lemma is the perfect tool when you have a space fully covered by a set of functional "patches" that are already defined, and you simply need to check if they stitch together nicely.

But what if you only have one patch? Suppose you have a function defined only on a closed subset $A$ of a space $X$, like a function defined only on the two points $\{-2, 2\}$ in the real line. You don't have another function on the complement to paste it with. Your goal is to extend the function to the entire space $X$ while preserving continuity. This is a different job entirely. It's not a matter of stitching, but of weaving new cloth to fill a hole. For this task, we need a different, more powerful tool called the ​​Tietze Extension Theorem​​, which guarantees that such an extension is always possible under certain reasonable conditions.

The Pasting Lemma, then, is our master tailor's rule for assembling a complete garment from a full set of pre-cut pieces. It ensures the integrity of the seams, guaranteeing a final product that is whole and continuous, a beautiful testament to the power of joining simple things together to create something complex and elegant.

After mastering the mechanics of a tool, the real joy comes from seeing what you can build with it. The Pasting Lemma, which we have just explored, is far more than a dry, technical result. It is a master artisan's secret, a fundamental principle of construction that allows us to take simple, well-behaved pieces and glue them together to form larger, more intricate, yet perfectly sound structures. In mathematics, this act of "gluing" functions continuously is the key that unlocks the door to algebraic topology, a field that seeks to understand the essential nature of shape by translating geometry into the language of algebra.

Let's embark on a journey to see how this humble lemma serves as the architectural foundation for some of the most beautiful and powerful ideas in modern mathematics.

Imagine you're tracing a route on a map. You have a path from your home to the library, and another from the library to the park. Intuitively, you can combine these to form a single, continuous journey from home to the park. In topology, these journeys are represented by continuous functions called paths. If we have a path $f$ from point $x$ to $y$, and a path $g$ from $y$ to $z$, we can define a new "concatenated" path, $h = f * g$, that takes us from $x$ to $z$.

The natural way to do this is to traverse the path $f$ at double speed for the first half of our journey (say, from time $t=0$ to $t=1/2$) and then traverse $g$ at double speed for the second half (from $t=1/2$ to $t=1$). This gives a formal definition:

But here we face a crucial question: is this new path $h$ truly continuous? The journey feels continuous, but the mathematical definition is split in two. What guarantees a seamless transition at the midpoint $t=1/2$? This is the first, and perhaps most fundamental, application of the Pasting Lemma. By defining our function on the two closed intervals $[0, 1/2]$ and $[1/2, 1]$, and by ensuring the pieces match up at their intersection—since $f(1) = y = g(0)$—the lemma assures us that the combined function $h$ is perfectly continuous.

This simple act of pasting paths has a profound consequence. It establishes that the relationship of "being connected by a path" is transitive. If there's a path from $x$ to $y$ and one from $y$ to $z$, then the existence of the continuous concatenated path means there is one from $x$ to $z$. When we add the obvious facts that any point is connected to itself (by a constant path) and that any path can be traversed in reverse, we discover that path-connectedness is an equivalence relation. This is a powerful organizing principle, as it allows us to partition any topological space into its "path components"—the distinct islands that one cannot travel between.

Furthermore, this building-block approach extends to spaces themselves. If we construct a space $X$ by joining two path-connected subspaces, $A$ and $B$, that share at least one point, is the entire space $X$ path-connected? Yes. To get from a point in $A$ to a point in $B$, we can simply travel within $A$ to a common point, and then travel from that point into $B$. This journey is, of course, a concatenated path whose continuity is once again guaranteed by our trusty lemma.

The Pasting Lemma's power is not confined to one-dimensional paths. It allows us to glue together higher-dimensional processes, most notably the continuous deformations known as homotopies. A homotopy is a way of continuously transforming one function into another, like smoothly morphing the letter 'C' into the letter 'I'.

Suppose we know how to deform a map $f$ into another map $g$, and we also know how to deform $g$ into a third map $h$. It stands to reason that we should be able to deform $f$ all the way to $h$. How? We simply perform the first deformation, and then immediately follow it with the second. To make this a single, unified deformation, we can run the first homotopy (from $f$ to $g$) during the time interval $t \in [0, 1/2]$ and the second homotopy (from $g$ to $h$) during $t \in [1/2, 1]$. This construction of a "concatenated homotopy" is perfectly analogous to path concatenation, and its continuity is, once again, a direct gift of the Pasting Lemma. This establishes that homotopy is a transitive relation.

This transitivity is the cornerstone of one of the most celebrated achievements of 20th-century mathematics: the fundamental group, $\pi_1(X, x_0)$. The elements of this group are not loops (paths starting and ending at the same point $x_0$), but rather homotopy classes of loops—families of loops that can be deformed into one another. The group operation is defined by path concatenation.

For this algebraic structure to be meaningful, the operation must respect the homotopy classes. That is, if we take two loops $f$ and $f'$ that are deformable into each other, and another pair $g$ and $g'$, their concatenations $f * g$ and $f' * g'$ must also be deformable into each other. Proving this is a beautiful geometric argument made rigorous by the Pasting Lemma. We are given the deformation from $f$ to $f'$ (a homotopy $F$) and the deformation from $g$ to $g'$ (a homotopy $G$). We can literally "paste" these two deformations side-by-side to create a larger, "quilted" deformation $H$ that smoothly transforms the path $f*g$ into $f'*g'$. The explicit construction stitches $F$ and $G$ together along the spatial parameter of the path:

This elegant construction, whose continuity is certified by the Pasting Lemma, is what ensures the group operation on $\pi_1(X, x_0)$ is well-defined. It is the crucial link that allows us to use the power and certainty of algebra to study the fluid and elusive world of shapes.

Our journey has shown the Pasting Lemma as a tool for creating new continuous functions. But we can shift our perspective and view these constructions themselves as functions between spaces of functions. For instance, path concatenation can be seen as a map that takes a pair of loops and outputs a single loop.

Here, $\Omega(X, x_0)$ is the "loop space" of $X$, the space of all possible loops based at $x_0$. A natural question arises: is this concatenation map itself continuous? In other words, if we make tiny, continuous changes to the input loops, does the resulting concatenated loop also change in a tiny, continuous way?

The answer is yes, and the proof is a wonderful, higher-level application of the Pasting Lemma. By analyzing the map in conjunction with the evaluation map (which simply evaluates a path at a given time), one can show that the entire process is continuous. The argument again involves splitting a domain into two closed pieces and applying the lemma, but this time the domain involves the abstract space of functions itself. This reveals that the structures we build are not just continuous, but the very act of building them is a continuous process.

This "pasting" principle is so fundamental that it echoes in even more abstract realms of topology. In advanced homotopy theory, certain maps called "cofibrations" are prized for their excellent behavior with respect to homotopies. A key theorem, which could be called a "pasting lemma for cofibrations," shows that if you glue spaces together in a particular way (a "pushout"), the property of being a cofibration is preserved. The proof strategy is a grand generalization of everything we have seen: one constructs pieces of a homotopy on different parts of a complex diagram and then uses the universal property of the pushout—a kind of abstract pasting lemma—to stitch them together into a single, coherent homotopy that proves the desired result.

From the simple act of joining two lines to verifying the algebraic foundations of topology and proving deep structural theorems, the Pasting Lemma is the silent, indispensable partner. It is a testament to the beauty of mathematics, where a single, clear idea can provide the logical thread that weaves together a vast and intricate tapestry of thought.