Image of a homomorphism

In the world of abstract algebra, a group homomorphism is a map between two groups that respects their underlying structure. But what does this 'respect for structure' truly mean? When we project one algebraic world onto another, what is the resulting shape? This projected form, known as the ​​image​​ of the homomorphism, is more than just a collection of elements; it is a 'shadow' that carries profound information about the original group. The central challenge this article addresses is understanding the nature of this shadow—its properties, its size, and its relationship to the group that casts it. By studying the image, we can decode the very essence of a homomorphism, revealing what is preserved and what is lost in translation. This article will guide you through this fundamental concept in two parts. First, in ​​Principles and Mechanisms​​, we will delve into the core definitions, explore how properties like cyclicity are preserved, and uncover the ultimate relationship between the domain, kernel, and image through the First Isomorphism Theorem. Following this, ​​Applications and Interdisciplinary Connections​​ will showcase how this seemingly abstract idea finds powerful expression in fields as diverse as physics, geometry, and topology, acting as a unifying bridge between different mathematical realms.

Imagine a powerful lamp casting a shadow of a complex three-dimensional sculpture onto a flat wall. The shadow is a two-dimensional projection. It might be smaller, and it has certainly lost a dimension of depth, but it's not just a random smudge. The shadow preserves the essential outline, the silhouette, of the original sculpture. A group homomorphism is much like this lamp. It's a map, a special kind of projection, from one group (our sculpture, the ​​domain​​) to another (the wall, the ​​codomain​​). The shadow it casts is called the ​​image​​. And just like a real shadow, the image tells us something profound about the original object, even if some information is lost in the projection.

At its core, the ​​image​​ of a homomorphism $\phi: G \to H$ is simply the set of all possible landing spots for elements of $G$ inside $H$. We denote it as $\text{Im}(\phi)$. So, $\text{Im}(\phi) = \{h \in H \mid h = \phi(g) \text{ for some } g \in G\}$.

Let's start with the simplest possible case. Consider the group of all integers under addition, $(\mathbb{Z}, +)$, as our domain. Let's map it to itself using the rule $\phi(n) = 6n$. What's the shadow, the image? For every integer $n$ we can think of, we get an output that is a multiple of 6. The set of all integers, an infinite line of equally spaced points, is mapped to a new line of points, but one where the gap between them has been stretched by a factor of six. The image is the set of all multiples of 6, which we write as $6\mathbb{Z}$. The image is a thinned-out version of the original, but it retains the essential structure of an infinite, ordered set of points.

What happens if the "wall" we project onto is finite? Let's take our same infinite group of integers, $\mathbb{Z}$, but this time project it onto the finite group of integers modulo 18, $\mathbb{Z}_{18}$. Suppose the homomorphism is defined by where it sends the number 1, say $\phi(1) = 6$. Because $\phi$ must preserve the group structure, the destination of any integer $n$ is fixed: $\phi(n) = n \times \phi(1) = 6n \pmod{18}$. The infinite line of multiples of 6 now gets "wrapped around" the clock face of $\mathbb{Z}_{18}$.

• $\phi(0) = 0$
• $\phi(1) = 6$
• $\phi(2) = 12$
• $\phi(3) = 18 \equiv 0 \pmod{18}$
• $\phi(4) = 24 \equiv 6 \pmod{18}$ ...and so on. The image doesn't grow infinitely; it traces out a repeating, finite pattern. The shadow cast by the infinite group is the small, finite set $\{0, 6, 12\}$.

This reveals a fundamental truth: the image of a homomorphism is not just some arbitrary collection of elements. It is always, without exception, a ​​subgroup​​ of the codomain $H$. Why must this be? The reason lies in the very definition of a homomorphism—it preserves the operation. If you take two elements in the image, say $h_1$ and $h_2$, they must have come from some elements in the domain, say $g_1$ and $g_2$. The product $h_1 h_2$ in the codomain is, by the homomorphism property, equal to $\phi(g_1) \phi(g_2) = \phi(g_1 g_2)$. Since $g_1 g_2$ is an element of the original group $G$, its image, $h_1 h_2$, must be in the image set. The shadow is structurally complete; it is a self-contained group in its own right.

A homomorphism is a structure-preserving map, so it's natural to ask what structural properties of the domain group $G$ are inherited by its image.

Perhaps the most important preservation property concerns cyclicity. If the domain group $G$ is ​​cyclic​​ (meaning it can be generated by a single element), then its image, $\text{Im}(\phi)$, must also be a cyclic group. This is a fantastically powerful constraint! If $G = \langle g \rangle$, then any element in $G$ is of the form $g^k$ for some integer $k$. The image of this element is $\phi(g^k) = (\phi(g))^k$. This means every single element in the image is just a power of the single element $\phi(g)$. The entire shadow is generated by the shadow of the original generator.

Imagine a homomorphism from the cyclic group $\mathbb{Z}_{12}$ to the dihedral group $D_6$, the symmetries of a hexagon. Let's say our map is defined by where it sends the generator $1 \in \mathbb{Z}_{12}$, for instance, $\phi(1) = r^2$, where $r$ is a rotation of the hexagon by 60 degrees. What is the image? We don't need to check all 12 elements. We know the image must be the cyclic subgroup generated by $\phi(1) = r^2$. The powers of $r^2$ are $\{(r^2)^0, (r^2)^1, (r^2)^2, \dots\} = \{e, r^2, r^4\}$. The image is the cyclic group of order 3, consisting only of rotations. Even though the codomain $D_6$ contains reflections, none of them appear in the image. The cyclic nature of the domain forces the image to be cyclic.

This leads to a beautiful generalization: if you have a homomorphism starting from a cyclic group like $\mathbb{Z}_{12}$, the only possible groups you can get as an image (up to isomorphism) are other cyclic groups whose order divides 12. You can get $\mathbb{Z}_6$, $\mathbb{Z}_4$, $\mathbb{Z}_3$, etc., but you could never get, for example, the Klein four-group ($\mathbb{Z}_2 \times \mathbb{Z}_2$), because it isn't cyclic. Similarly, if the domain is ​​abelian​​ (commutative), its image must also be abelian. The shadow cannot be more complex in its commutation relations than the object that casts it.

We've seen that the image is a simplified version of the domain. But how much is lost? Is there a precise mathematical law governing this relationship? The answer is yes, and it is one of the pillars of modern algebra: the ​​First Isomorphism Theorem​​.

This theorem connects the image to another crucial concept: the ​​kernel​​ of the homomorphism. The kernel, denoted $\ker(\phi)$, is the set of all elements in the domain $G$ that are "crushed" or "annihilated" by the map, sending them to the identity element $e_H$ in the codomain $H$. It's the part of the sculpture that is directly in line with the lamp, leaving no shadow.

The First Isomorphism Theorem states that the image of the homomorphism is structurally identical (isomorphic) to the domain group after you "collapse" all the information contained in the kernel. This "collapsed" or "quotient" group is written as $G/\ker(\phi)$. In essence: $\text{Im}(\phi) \cong G/\ker(\phi)$ The image's structure is precisely the domain's structure, modulo the kernel.

An immediate and incredibly useful consequence of this theorem concerns the sizes of finite groups. The order of a quotient group $|G/N|$ is simply $\frac{|G|}{|N|}$. Applying this to our theorem gives the fundamental counting formula: $|\text{Im}(\phi)| = \frac{|G|}{|\ker(\phi)|}$ The size of the shadow is the size of the original object divided by the size of the part that got lost. This relationship is universal.

For example, if you have a homomorphism from a cyclic group of order 28, $C_{28}$, and you know that its kernel has 7 elements, you don't need to know anything else about the map or the codomain. The image must have an order of $\frac{28}{7} = 4$. This works for any group, not just cyclic ones. A homomorphism from the non-abelian dihedral group $D_6$ (order 12) with a kernel of order 3 will invariably produce an image of order $\frac{12}{3} = 4$. This theorem provides a beautiful, unifying bridge between the three fundamental concepts: the domain, the kernel, and the image.

With these tools, we can become architects of abstract algebra, reasoning not just about a single homomorphism but about the entire universe of possibilities. What kinds of images can exist for a map between two given groups, $G$ and $H$?

1. From our shadow analogy, the image $\text{Im}(\phi)$ is a subgroup of the "wall" $H$. By Lagrange's Theorem, a foundational result in group theory, the order of any subgroup must divide the order of the group. So, ​​$|\text{Im}(\phi)|$ must be a divisor of $|H|$​​.
2. From the First Isomorphism Theorem, $|\text{Im}(\phi)| = |G|/|\ker(\phi)|$. This means ​​$|\text{Im}(\phi)|$ must be a divisor of $|G|$​​.

Putting these two constraints together gives us a powerful predictive tool: the order of any possible image must divide both $|G|$ and $|H|$. In other words, the possible orders for an image are limited to the common divisors of the orders of the domain and codomain. $|\text{Im}(\phi)| \text{ must divide } \gcd(|G|, |H|)$ For any homomorphism $\phi: \mathbb{Z}_{36} \to \mathbb{Z}_{48}$, we know instantly that the order of the image must be a divisor of $\gcd(36, 48) = 12$. The possible orders are thus restricted to the divisors of 12. In fact, one can show that for any non-trivial map, all divisors of 12 other than 1 are possible image sizes: 2, 3, 4, 6, and 12.

But there's one final, subtle constraint: the inherent structure of the domain itself. Consider the alternating group $A_5$, the group of even permutations on 5 elements. This group is famous for being ​​simple​​, which means its only normal subgroups are the trivial group and itself. It is also non-abelian. What happens if we try to map $A_5$ to an abelian group $H$, like the integers or $\mathbb{Z}_n$? The image must be a subgroup of $H$, and therefore must be abelian. By the First Isomorphism Theorem, $A_5/\ker(\phi)$ must be abelian. The kernel is a normal subgroup, so for $A_5$, $\ker(\phi)$ must be either the trivial group $\{e\}$ or all of $A_5$.

• If $\ker(\phi) = \{e\}$, then the image would be isomorphic to $A_5$ itself, implying $A_5$ is abelian. This is false.
• Therefore, the only possibility is that $\ker(\phi) = A_5$. Every single element of $A_5$ is crushed down to the identity.

This means the image is the trivial group $\{e\}$. The group $A_5$ is so fundamentally non-abelian that it refuses to cast any non-trivial shadow on an abelian wall. Any attempt to do so collapses the entire structure into a single, infinitesimal point. The image, our shadow, reveals not only what is preserved, but also what is so essential to a group's identity that it cannot be projected away without destroying the entire structure.

We have spent our time looking at the machinery of homomorphisms, seeing how they are the proper way to discuss mappings between algebraic structures. We have defined the image as the set of all landing spots for such a map. On the face of it, this might seem like a mere terminological convenience, a way to name a particular subset of the codomain. But to leave it there would be like describing a telescope as "a tube with glass in it." The real magic is not in what it is, but in what it lets us see.

The image of a homomorphism is not just a dusty subset; it is the shadow that one mathematical world casts upon another. By studying the shape and size of this shadow, we can learn an astonishing amount about both worlds—the one casting the shadow and the one it falls upon. In this chapter, we will zoom out from the formal definitions and embark on a journey to see what this concept of an "image" actually does out in the wild. We shall find it acting as a powerful tool for classification, a fundamental constraint in physics, a bridge between algebra and geometry, and a subtle probe into the very shape of space.

Let's begin with something familiar: shuffling a deck of cards. The set of all possible shuffles of $n$ items forms a group, the symmetric group $S_n$. This group can feel like a chaotic mess of possibilities. Yet, there is a wonderfully simple way to bring order to it. We can define a homomorphism, the sign map, which sends every permutation to one of two numbers: $+1$ or $-1$. The map assigns $+1$ if the permutation can be achieved by an even number of two-card swaps (transpositions), and $-1$ if it requires an odd number.

What is the image of this map? For any group of shuffles involving two or more cards, the image is always the full set $\{-1, 1\}$. Why? Because the "do nothing" shuffle is even (0 swaps), so its image is $+1$. And a single swap of two cards is odd, so its image is $-1$. Since we land on both possible values, the image is the entire codomain. This simple fact is profound. The image tells us that the world of permutations is cleanly and perfectly divided into two non-overlapping halves: the "even" ones and the "odd" ones. This classification is the bedrock upon which the theory of determinants is built, and it even finds its way into quantum mechanics, distinguishing the behavior of fundamental particles.

This idea of an image acting as a constraint becomes even clearer in physics. The state of a quantum system is described by a vector, and its evolution over time—how it changes from one moment to the next—is described by multiplication by a special kind of matrix from the unitary group, $U(n)$. These matrices have the crucial property that they preserve the length of vectors, which corresponds to the physical law that total probability must always be 1. Now, let’s consider a homomorphism that is immensely useful for matrices: the determinant. This map takes an $n \times n$ matrix and crunches it down to a single complex number. What is the image of the determinant map when we apply it only to the unitary matrices?

One might guess the image is all non-zero complex numbers, $\mathbb{C}^*$. But it is not. The condition of being unitary—of preserving length—places a powerful constraint on the determinant. If a matrix $A$ is in $U(n)$, then its determinant must have an absolute value of 1. This means the image is not the entire complex plane, but is confined to the unit circle, the group known as $U(1)$. The image reveals a deep physical principle: a quantum evolution, while it can rotate the "phase" of a state (which corresponds to moving around the circle $U(1)$), can never change its overall magnitude. The image of the homomorphism is the mathematical embodiment of a physical conservation law.

One of the most beautiful themes in modern mathematics is the intimate dance between algebra and geometry. On the surface, they seem like different subjects: one about symbols and equations, the other about shapes and spaces. The image of a homomorphism often serves as the bridge between them.

Consider the ring of all polynomials with real coefficients, $\mathbb{R}[x]$. This is an algebraic world. Now, let’s define a homomorphism on it. We'll map a polynomial $p(x)$ to a new polynomial, $(x^2 - 9)p(x)$. What is the image of this map? It is the set of all polynomials that can be written in the form $(x^2 - 9)p(x)$ for some $p(x)$. Phrased algebraically, it is the ideal generated by $x^2-9$. But what is this set geometrically? A polynomial is in this image if and only if it has $(x-3)$ and $(x+3)$ as factors. This is the same as saying it must be equal to zero when you plug in $x=3$ or $x=-3$. Look at what happened! Our purely algebraic object—the image of a homomorphism—is perfectly described by a geometric condition: the set of all polynomials that vanish on the set of points $\{-3, 3\}$. This is a foundational example of the "algebra-geometry dictionary," where ideals in a ring correspond to geometric shapes. The image is the translation key.

This connection becomes even more visual and dynamic when we look at Lie groups. Imagine a doughnut, or what a mathematician calls a 2-torus, $T^2$. We can think of it as a product of two circles, $U(1) \times U(1)$. Now, imagine a straight, infinite line, which is the group of real numbers $\mathbb{R}$. We can define a homomorphism from the line to the torus that takes a number $t$ and maps it to a point on the torus by winding around the first circle $\alpha t$ times and the second circle $\beta t$ times. The path traced out by this mapping is the image.

What does this path look like? Does it repeat itself, forming a closed loop? Or does it wander around forever, never quite closing? Astonishingly, the answer depends on a simple algebraic property of the numbers $\alpha$ and $\beta$. If their ratio $\alpha/\beta$ is a rational number, the path will eventually meet up with itself and form a closed loop. But if the ratio $\alpha/\beta$ is irrational, the path will wind around the torus forever, never intersecting itself, and getting arbitrarily close to every single point on the torus. In this case, we say the image is a dense subgroup. This is a magnificent result! A subtle number-theoretic property of two constants completely determines the global, geometric, and topological character of the image. This single idea connects group theory to dynamical systems and the study of quasi-periodic motion.

Perhaps the most surprising and powerful applications of the homomorphism's image are found in algebraic topology, a field dedicated to studying the essential properties of shapes by assigning algebraic objects to them.

To get our bearings, let's start with the simplest possible continuous map between two topological spaces, $X$ and $Y$: a constant map that takes every point in $X$ and sends it to a single point $y_0$ in $Y$. This geometric action of "crushing $X$ to a point" has an algebraic counterpart. It induces a homomorphism on the fundamental groups (or groupoids), which are algebraic invariants that encode information about the loops in a space. What is the image of this induced homomorphism? It is the trivial group, containing only the identity element. All the rich loop structure of $X$ is mapped to the "do nothing" loop at $y_0$. This is a crucial sanity check. Our algebraic tools are behaving exactly as our geometric intuition would demand. If you geometrically erase all information, the algebraic picture should become trivial as well.

Now for something more interesting. The circle, $S^1$, is a fundamental object in topology. Its fundamental group, $\pi_1(S^1)$, is the group of integers, $\mathbb{Z}$, where each integer corresponds to how many times you wind around the circle. The group $\mathbb{Z}$ is the quintessential cyclic group, generated by a single element, the number 1. What happens when we map a circle into some other, more complicated space $Y$? This map induces a homomorphism $f_*: \pi_1(S^1) \to \pi_1(Y)$. A basic theorem of group theory states that the image of a homomorphism from a cyclic group must itself be a cyclic group. Therefore, the image, $\text{Im}(f_*)$, is always a cyclic subgroup of $\pi_1(Y)$. This means that by mapping a circle into a space, we are using it as a probe. We are "plucking" the space and listening to the note it produces. The image of the map is that note, an echo which tells us about the cyclic structures hidden deep within the fabric of the space $Y$.

This principle finds its grandest expression in the theory of covering spaces. A covering space $E$ of a base space $B$ is an "unwrapped" version of $B$, like the infinite line $\mathbb{R}$ is an unwrapped version of the circle $S^1$. A central question is: if we have a map from some space $Y$ into the base $B$, can we "lift" it to a map from $Y$ into the covering space $E$? The answer is given by the lifting criterion, a theorem that rests entirely on comparing the images of two homomorphisms. The lift exists if and only if the image of the map induced by $f: Y \to B$ is a subgroup of the image of the map induced by the covering projection $p: E \to B$. In symbols, $\text{Im}(f_*) \subseteq \text{Im}(p_*)$.

It’s like asking for permission. The group $\text{Im}(p_*)$ represents all the loop structures in the base $B$ that "come from" the covering space $E$. The group $\text{Im}(f_*)$ represents the loop structures that the map $f$ is trying to "draw" in $B$. A lift is possible only if the picture you want to draw is made from elements that are available from the covering space. Once again, the comparison of two images, two shadows cast in the world of groups, provides the definitive answer to a deep geometric question.

This theme continues into even more advanced theories. In homology theory, a "long exact sequence" provides a series of connected homomorphisms. The image of one map becomes the kernel of the next, weaving a beautiful algebraic tapestry. For instance, the image of the connecting homomorphism $\partial_*: H_2(X, A) \to H_1(A)$ tells us exactly which loops on the boundary surface $A$ of a 3-dimensional object $X$ are actually the edges of surfaces lying inside $X$. For a "handlebody" with $g$ holes, the rank of this image is precisely $g$, wonderfully tying the algebraic size of the image to the geometric complexity of the object.

From the simple sorting of shuffles to the abstract shape of unseen dimensions, the image of a homomorphism is a recurring character in the story of science. It is the footprint left by one structure upon another, and by studying its shape, its size, and its properties, we unlock the secrets of them all.