Attaching Maps

In the abstract world of topology, how are complex, intricate shapes constructed from the simplest possible ingredients? If we think of points, lines, and disks as our basic building blocks, we are faced with a fundamental question: what instructions guide their assembly into structures like a torus, a Klein bottle, or even more exotic forms? This process is not arbitrary; it is governed by a precise and powerful concept known as the ​​attaching map​​, which serves as the architectural blueprint for gluing new pieces onto an existing space. This article demystifies this "art of gluing," revealing it as the key mechanism that dictates the final form and properties of a topological space.

This article will guide you through the theory and application of this foundational concept. The first chapter, ​​"Principles and Mechanisms,"​​ lays the groundwork by defining what an attaching map is, explaining the crucial rule of gluing along boundaries, and demonstrating how different maps—such as those with varying "degrees"—can produce dramatically different spaces from the same raw materials. Following this, the chapter on ​​"Applications and Interdisciplinary Connections"​​ showcases the immense power of this technique, exploring how to engineer spaces with desired algebraic features and revealing how famous objects from geometry and physics are built using these very blueprints.

Imagine you are a sculptor, but instead of clay or marble, your materials are pure, idealized shapes: points, lines, disks, and spheres. How would you assemble these simple building blocks into the vast and intricate sculptures that we call topological spaces—like a torus, a projective plane, or something far more exotic? The secret lies not just in the pieces themselves, but in the instructions for how they are glued together. This set of instructions, this art of gluing, is what topologists call an ​​attaching map​​. It's the genetic code that determines the final form and properties of the universe you are building.

Let's start with the basics. Our building blocks are called ​​cells​​. A 0-cell is a point, a 1-cell is an open line segment, a 2-cell is an open disk, and an $n$-cell is an open $n$-dimensional ball. To build a space, we start with a collection of points (the 0-skeleton) and then attach higher-dimensional cells one by one.

But how, exactly, do we "attach" a cell? Suppose we want to attach a 2-cell (a disk, $D^2$) to a space we've already built, say, a circle $S^1$. A natural but incorrect impulse might be to somehow merge the inside of the disk with the circle. This, however, is not how the game is played. The fundamental rule of attachment is that you only glue along the boundary of the new cell. The interior of the cell remains pristine and new, forming a new open region in our space.

So, to attach our 2-disk $D^2$, we must provide instructions on how to glue its boundary, which is a circle $S^1$, onto the existing space. The domain of our attaching map must be the boundary of the cell we are adding. This map, $\phi: S^1 \to \text{existing space}$, is the blueprint. For every point on the disk's boundary, the map tells us which point in the existing space it gets glued to.

Let's make this concrete with a famous example: the torus, or the surface of a donut. A very efficient way to build a torus is to start with one point (a 0-cell). Then, we attach two 1-cells, say $a$ and $b$. We attach each 1-cell (a line segment) by gluing both of its endpoints to our single point. The result is two circles joined at a point, a shape known as the wedge sum of two circles, $S^1 \vee S^1$. This is our 1-skeleton.

Now for the magic. We take a single 2-cell, a flat disk which you can imagine as a square sheet of rubber, and we need to glue its boundary to our figure-eight skeleton. The boundary of the square has four sides. We traverse the first side and glue it along loop $a$, the second side along loop $b$, the third side back along loop $a$ in the reverse direction, and the final side back along loop $b$ in reverse. The attaching map takes the boundary circle of the disk and traces the path $aba^{-1}b^{-1}$ on the $S^1 \vee S^1$ skeleton. The result? A perfect torus. The attaching map's domain was the boundary of our 2-cell, a circle, and its codomain was the 1-skeleton we had already built, $S^1 \vee S^1$.

This leads us to the most beautiful and profound idea in this construction: the final shape of the space is determined almost entirely by the attaching map. The cells are generic, but the gluing instructions are everything.

Let's consider a simpler scenario. We start with a single 0-cell and attach a single 1-cell to it to form a circle, $S^1$. Now, we want to attach a 2-cell (a disk) to this circle. Our attaching map $\phi$ will be a map from the boundary of the disk (a circle) to the circle we just made. How many ways can you map a circle to a circle?

Intuitively, you can just lay it on top, a one-to-one correspondence. Or you could wrap it around twice. Or three times. Or seven times, perhaps in the opposite direction. This "number of wraps" is a well-defined integer called the ​​degree​​ of the map. For a map from the circle (represented as unit complex numbers $\{z \in \mathbb{C} : |z|=1\}$) to itself, the function $\phi(z) = z^n$ has degree $n$. A positive $n$ means it wraps $n$ times in the same direction, a negative $n$ means it wraps $|n|$ times in the reverse direction, and $n=0$ means it maps the whole boundary to a single point.

Let's see what happens when we choose different degrees for our attaching map $\phi: \partial D^2 \to S^1$.

• ​​Degree 2:​​ We attach the disk by wrapping its boundary twice around the circle. The resulting space is none other than the ​​real projective plane​​, $\mathbb{R}P^2$—a bizarre one-sided surface where if you walk in a straight line, you end up back where you started but mirror-reversed. Algebraically, we can see this because the degree of the attaching map becomes the key number in calculating properties like homology. For a degree $k$ attachment to a circle, the first homology group $H_1$ of the new space is $\mathbb{Z}/k\mathbb{Z}$. Since we know $H_1(\mathbb{R}P^2) \cong \mathbb{Z}/2\mathbb{Z}$, the attaching map must have an absolute degree of 2. Geometrically, you can picture $\mathbb{R}P^2$ as a disk where you identify opposite points on its boundary. If you trace your finger along the entire boundary of this disk, its image in the final space is a circle that you've traversed twice. This gives a beautiful visual for the degree-2 attachment.

The truly wonderful insight here comes from thinking about loops. The fundamental group, $\pi_1$, of our initial circle is $\mathbb{Z}$, generated by a loop $a$ that goes around once. When we attach the 2-cell with a degree-2 map, we are gluing the disk's boundary along the path corresponding to $a^2$ (going around twice). In the new space, this path, $a^2$, is now the edge of a surface—the disk we just attached! Since it bounds a surface, we can continuously shrink this loop across the disk down to a single point. It has become trivial, or "nullhomotopic". This is the mechanism: attaching a cell forces the attaching loop to become trivial. Thus, in our new space's fundamental group, we have the new relation $a^2=1$, which gives us $\pi_1(\mathbb{R}P^2) \cong \mathbb{Z}/2\mathbb{Z}$.

This principle is a recipe for creation. By simply changing the degree of the attaching map, we can manufacture entirely different universes.

• ​​Degree 1:​​ What if we attach our disk with a degree-1 map? We're gluing the disk's boundary to the circle with a simple one-to-one mapping. This is like putting a lid on a can. The original hole of the circle is now filled in. The resulting space is just a disk, which can be shrunk to a point. The original circle, as a subspace of this new space, is now "nullhomotopic"—it can be shrunk to a point. It turns out this only happens if the degree of the attachment is $\pm 1$. For any other degree, the hole isn't truly filled, and the original circle remains "trapped".

• ​​Degree 3, 4, 5...:​​ What if we attach with degree 3? The resulting space has a fundamental group of $\mathbb{Z}/3\mathbb{Z}$ and a first homology group of $\mathbb{Z}/3\mathbb{Z}$. This space is not a sphere, not a torus, and not a projective plane. It is a different kind of object entirely, sometimes called a Moore space. We have the power to create a space with any desired finite cyclic homology group $\mathbb{Z}/n\mathbb{Z}$ simply by choosing our attaching map to have degree $n$. We are topological engineers!

• ​​More Complex Recipes:​​ We can apply this principle to more complicated skeletons and more complicated attaching paths. Let's revisit the torus. We started with a figure-eight space, $S^1 \vee S^1$, whose fundamental group is the free group on two generators, $\langle a,b \rangle$. In this group, $a$ and $b$ do not commute; $ab \neq ba$. The path corresponding to the commutator, $aba^{-1}b^{-1}$, is a non-trivial loop. To get the torus, whose fundamental group is the commutative group $\mathbb{Z}^2$, we need to force $a$ and $b$ to commute, which is equivalent to enforcing the relation $aba^{-1}b^{-1}=1$. How do we do it? Simple! We attach a 2-cell with its boundary tracing the path $aba^{-1}b^{-1}$. This loop is now "killed", forcing the relation and yielding the correct fundamental group for the torus.

This incredible power comes with a few rules. One of the most important is a "tidiness" condition known as ​​closure-finiteness​​. It states that the closure of any given cell can only intersect a finite number of other cells. This rule prevents us from building certain pathological spaces, like trying to glue a single disk onto an infinite line of 1-cells in such a way that its boundary comes arbitrarily close to all of them. It ensures that our constructions are, in a sense, "locally tame".

This entire framework, of cells and attaching maps, is not confined to the simple, low-dimensional worlds we've explored. The principle scales up to dizzying heights of abstraction. Imagine attaching a 7-dimensional cell to a 3-dimensional sphere. The attaching map would be a function from the boundary of the 7-cell (a 6-sphere, $S^6$) to the 3-sphere. The "degree" of this map is no longer a simple integer but an element in a far more mysterious object called the sixth ​​homotopy group of the 3-sphere​​, $\pi_6(S^3)$, which happens to be a cyclic group of order 12, $\mathbb{Z}/12\mathbb{Z}$.

Even in this strange world, the principle holds: the homotopy class of the attaching map determines the homotopy type of the resulting space. We can ask questions like: how many different kinds of spaces can we build if our attaching map corresponds to an element of order 4 in this group? By analyzing the group structure, we find there are two such elements, 3 and 9. But because reversing the orientation of the attachment ($\pm f$) doesn't change the space type, these two maps are equivalent. The astonishing conclusion is that there is only one such space possible. From wrapping a string around a peg to classifying 7-dimensional attachments to a 3-sphere, the fundamental idea remains the same. The beauty of the attaching map is its power to translate a simple, intuitive act of gluing into a precise and powerful engine for creating and understanding the boundless universe of topological shapes.

After exploring the foundational principles of attaching maps, we now arrive at a thrilling vista. We are like architects who have just mastered the properties of bricks and mortar. Now, we can ask: What can we build? This is where the true power and beauty of the concept shine, revealing profound connections across mathematics and even into physics. The attaching map is not merely a technical device; it is the blueprint, the genetic code, that dictates the form and function of a topological space. We are about to embark on a journey of "topological engineering," where we construct spaces with bespoke properties and, in doing so, discover that some of the most famous structures in geometry are built from these very plans.

The most intuitive property of a space is its "loopiness"—the different ways one can journey through it and return to the start. The fundamental group, $\pi_1(X)$, is the rigorous language for this concept. With attaching maps, we can become masters of this language, sculpting spaces with precisely the fundamental groups we desire.

Imagine you have a circle, $S^1$, whose fundamental group is the integers, $\mathbb{Z}$, corresponding to how many times a loop winds around it. What if you want to create a space where looping around, say, $n$ times is equivalent to not moving at all? You can achieve this by taking a 2-dimensional disk, $e^2$, and gluing its boundary circle onto your original circle $S^1$. But how you glue it is everything. If you trace the boundary of the disk around the circle $n$ times before sealing it shut, you've essentially provided an escape route. Any loop that wraps $n$ times can now be "shrunk" across this attached disk, making it trivial. The result? The new fundamental group is precisely the cyclic group of order $n$, $\mathbb{Z}/n\mathbb{Z}$. The degree of the attaching map directly engineers the torsion in the fundamental group.

The recipes can be far more intricate. Consider starting not with a circle, but with a figure-eight, $S^1 \vee S^1$, which has two fundamental loops, let's call them $a$ and $b$. To construct the famous and perplexing Klein bottle, one attaches a 2-cell (a square) according to a very specific set of instructions for its four edges. Traversing the boundary spells out the word $aba^{-1}b$. This "word" is the attaching map's signature in the fundamental group of the figure-eight. By gluing the disk along this path, we impose the relation $aba^{-1}b=1$ on our space. This single algebraic relation, born from the attaching map, captures the entire geometric essence of the Klein bottle—its one-sidedness and its peculiar twist.

What happens if we give a space conflicting instructions? Suppose we attach one disk to a circle by wrapping it 4 times ($a^4=1$) and another disk by wrapping it 6 times ($a^6=1$). A loop in this new space must obey both rules. Just as in elementary number theory, a number divisible by both 4 and 6 must be divisible by their least common multiple, a loop trivialized by both relations must be a multiple of the "strongest common relation." Algebraically, this corresponds to the greatest common divisor, $\gcd(4,6)=2$. The resulting space, despite the more complex instructions, has a simple fundamental group of $\mathbb{Z}/2\mathbb{Z}$. The attaching maps compete, and the resulting structure is a compromise dictated by pure arithmetic.

The influence of an attaching map runs much deeper than the fundamental group. It leaves its fingerprints on other algebraic invariants like homology and homotopy groups, which probe a space's structure in higher dimensions.

Let's revisit our space $X_n$ made by attaching a 2-cell to a circle with a degree $n$ map. We saw it shaped the fundamental group. But if we examine its cohomology—a sort of dual perspective on its structure—we find another echo of our construction. The second cohomology group, $\tilde{H}^2(X_n; \mathbb{Z})$, turns out to be exactly $\mathbb{Z}/n\mathbb{Z}$. The same is true for homology. The integer $n$ from our attaching map reappears, a testament to the consistency and interconnectedness of these algebraic probes. The geometric act of gluing has simultaneous and harmonious consequences in both homotopy and homology.

This principle is not confined to one or two dimensions. Imagine attaching a 3-dimensional cell to a 2-sphere, $S^2$. The attaching map is a map from a sphere to a sphere, $\phi: S^2 \to S^2$, which itself is characterized by an integer degree. If we use a map of degree $k$, we are imposing a relation on the second homotopy group, $\pi_2(S^2)$. The resulting space $X$ will have a second homotopy group of $\pi_2(X) \cong \mathbb{Z}/k\mathbb{Z}$. The general principle is magnificent in its simplicity: attaching an $(n+1)$-cell to a space $A$ via a map $\phi: S^n \to A$ introduces a relation into the $n$-th homotopy group, $\pi_n(A)$, dictated by the class of $\phi$.

This method of construction is no mere curiosity for creating abstract examples. It is the very method by which nature and mathematics build some of their most important and beautiful structures.

Consider the real projective plane, $\mathbb{R}P^2$, the space of all lines through the origin in 3D space. It is a fundamental object in geometry. If we study this space using the tools of calculus, specifically Morse theory, we find that any "nice" function on $\mathbb{R}P^2$ naturally decomposes it into a minimal set of cells: one each in dimensions 0, 1, and 2. The 1-skeleton is a circle. How must the 2-cell be attached to create $\mathbb{R}P^2$? The deep topological properties of $\mathbb{R}P^2$—specifically, that its first homology group is $\mathbb{Z}/2\mathbb{Z}$—force the answer upon us. The attaching map must have degree 2. The global topology of the manifold dictates its local construction blueprint.

The story becomes even more spectacular with the complex projective plane, $\mathbb{C}P^2$. This space is central to quantum mechanics and algebraic geometry. Its cell structure is simple: one cell each in dimensions 0, 2, and 4. This means we build it by attaching a 4-cell to a 2-sphere. The attaching map is a map $\phi: S^3 \to S^2$. Of the infinitely many possible such maps, which one builds $\mathbb{C}P^2$? The answer is astonishing: it is the celebrated Hopf map, a generator of the homotopy group $\pi_3(S^2)$. This is no coincidence. The rich algebraic structure of $\mathbb{C}P^2$, particularly its non-trivial "cup product" multiplication in cohomology, is a direct consequence of the non-triviality of the Hopf map as the gluing instruction.

Even a seemingly simple space like the product of two spheres, $S^p \times S^q$, reveals a secret when viewed through this lens. Its cell structure requires a $(p+q)$-cell to be attached to the wedge $S^p \vee S^q$. The attaching map is not trivial; it is a highly non-trivial homotopy element known as the Whitehead product of the two spheres. This sophisticated attaching map is precisely what's needed to give the cohomology of $S^p \times S^q$ its familiar product structure.

These constructions provide a fertile playground for exploring deeper phenomena. What happens if we take a space like $\mathbb{C}P^2$ (or its cousins $X_k$ from problem and apply a standard operation like suspension? Suspending a space is like spinning it in a higher dimension. The cell structure transforms in a predictable way: an $n$-cell becomes an $(n+1)$-cell, and the attaching maps are likewise suspended. The Hopf map $\eta: S^3 \to S^2$ becomes a map $\Sigma \eta: S^4 \to S^3$. The properties of this new map are a central question in homotopy theory. It turns out that the suspended map $k \cdot \Sigma \eta$ is trivial if and only if $k$ is an even number. This subtle arithmetic property, revealed in the simple laboratory of a cell complex, is a window into the fantastically complex structure of the homotopy groups of spheres.

From twisting loops to forging the arenas of modern geometry, the principle of the attaching map stands as a testament to a profound unity in mathematics. It is the bridge between the intuitive, geometric act of gluing and the powerful, abstract machinery of algebra. Each attaching map is a musical note in a grand symphony, defining the harmony and structure of the resulting space. By learning to write this music, we have gained the ability not just to analyze shapes, but to create them, revealing an intricate and beautiful order hidden within the world of topology.