Conjugation and the Product of Subgroups
Do or do not. There is no try. — Yoda.
Introduction
In group theory, we often want to know if two elements (or two subgroups) are structurally “the same,” just viewed from a different perspective. We also want to know how to combine two small subgroups to build a larger one.
These two goals lead us to the concepts of Conjugation and the Product of Subgroups.
Conjugate Elements and Subgroups
Definition: Let $G$ be an arbitrary group, and let $g_1, g_2 \in G$. We say that $g_1$ is a conjugate of $g_2$ if there exists some element $a \in G$ such that $g_1 = a g_2 a^{-1}$.
Geometric Intuition (Change of Perspective): What does $a g_2 a^{-1}$ actually mean? Imagine $g_2$ is an action, like “turn left.” Imagine $a$ is a translation, like “walk to the other side of the room.” The conjugate action ($a g_2 a^{-1}$) means:
- Reverse your translation ($a^{-1}$: walk back across the room).
- Perform the action ($g_2$: turn left).
- Re-apply the translation ($a$: walk back to your spot). You just performed a “turn left” from a completely different perspective! Conjugation is the algebraic equivalent of a Change of Basis in linear algebra.
Conjugation in Abelian Groups
If a group is strictly commutative (Abelian), the concept of conjugation collapses completely.
Proposition: Let $G$ be an Abelian group. Two elements $g_1, g_2 \in G$ are conjugates if and only if they are the exact same element ($g_1 = g_2$).
Proof:
Suppose $g_1$ is a conjugate of $g_2$. By definition, there exists some $a \in G$ such that: $g_1 = a g_2 a^{-1}$
Because the group is Abelian, we can commute $g_2$ and $a$: $g_1 = (g_2 a) a^{-1}$
By associativity and the definition of an inverse: $g_1 = g_2 (a a^{-1}) = g_2 (e) = g_2$. Thus, an element is only conjugate to itself. $\blacksquare$
Preserving Algebraic Structure
Because conjugation is just a change of perspective, it preserves the fundamental algebraic structure of the element.
Proposition: If $g_1$ is a conjugate of $g_2$, they have the exact same order.
Proof:
Let $g_1 = a g_2 a^{-1}$. Suppose the order of $g_2$ is $n$, meaning $g_2^n = e$.
Let’s raise $g_1$ to the $n$-th power: $(g_1)^n = (a g_2 a^{-1})^n = (a g_2 a^{-1})(a g_2 a^{-1}) \dots (a g_2 a^{-1})$
Notice what happens in the middle! Every $a^{-1}$ perfectly cancels out with the adjacent $a$: $(g_1)^n = a (g_2 g_2 \dots g_2) a^{-1} = a (g_2^n) a^{-1}$
Because $g_2^n = e$: $(g_1)^n = a (e) a^{-1} = a a^{-1} = e$. Because they both collapse to the identity at exactly the power of $n$, their orders are identical. $\blacksquare$
Conjugate Subgroups
We can apply this exact same “change of perspective” to an entire subgroup.
Definition: Let $H$ be a subgroup of $G$. The conjugate of $H$ by an element $g \in G$ is the set $g H g^{-1} = \{g h g^{-1} \mid h \in H\}$.
Proposition: The conjugate set $g H g^{-1}$ is always a valid subgroup of $G$.
Proof (Using the One-Step Subgroup Test):
Let $x, y \in g H g^{-1}$. By definition, $x = g h_1 g^{-1}$ and $y = g h_2 g^{-1}$ for some $h_1, h_2 \in H$. We must prove that $x y^{-1}$ is in the conjugate set.
$x y^{-1} = (g h_1 g^{-1})(g h_2 g^{-1})^{-1}$
Apply the “Shoes and Socks” principle to invert the second term: $x y^{-1} = (g h_1 g^{-1}) (g h_2^{-1} g^{-1})$
The middle terms cancel ($g^{-1} g = e$): $x y^{-1} = g (h_1 h_2^{-1}) g^{-1}$
Because $H$ is a valid subgroup, it is closed under the One-Step test, meaning $(h_1 h_2^{-1}) \in H$. Therefore, our final result is exactly in the form $g (\text{an element of } H) g^{-1}$, which means it belongs to $g H g^{-1}$. $\blacksquare$
Note: Just because $g H g^{-1}$ is a subgroup does NOT mean it equals $H$. If $g H g^{-1} = H$ for every $g \in G$, we call $H$ a Normal Subgroup—a crucial concept for later!
Proposition. Let $G$ be an Abelian group and $H_1 \le G, H_2 \le G$, then $H_1$ is a conjugate of $H_2$ if and only if $H_1 = H_2$.
Proof.
$\rightarrow$ Suppose $H_1$ is a conjugate of $H_2 \implies \exists a \in G: H_1 = aH_2a^{-1} \implies H_1a = aH_2 \implies$ [$G$ is Abelian] $aH_1 = aH_2 \implies$ [Multiplying both sides by $a^{-1}$] $H_1 = H_2$.
$\leftarrow$ Suppose $H_1 = H_2$. Let $e \in G$ be the identity element of $G$. Then, $H_1 = eH_2e^{-1}$ $\blacksquare$
The Product of Subgroups
If you have two subgroups, $H$ and $K$, you might want to combine them to form a larger subgroup. You cannot simply take their union ($H \cup K$). The union of two subgroups is usually not a subgroup because it is generally not closed under the group operation.
For example, consider the additive group $(\mathbb{Z},+)$: $H = 2\mathbb{Z} = \{\ldots,-4,-2,0,2,4,\ldots\}, K = 3\mathbb{Z} = \{\ldots,-6,-3,0,3,6,\ldots\}.$
Both are subgroups of $\mathbb{Z}$. Their union is $H \cup K = \{\ldots,-6,-4,-3,-2,0,2,3,4,6,\ldots\}.$ But $2 \in H,\qquad 3 \in K,$ while $2+3=5 \notin H \cup K.$
Instead, we define the Product of Subgroups.
Definition: Let $H$ and $K$ be subgroups of a group $G$. The product $HK$ is the set of all possible combinations generated by multiplying an element of $H$ by an element of $K$. $HK = \{hk \mid h \in H, k \in K\}$
When is $HK$ a Subgroup? Is the product of two subgroups automatically a subgroup? No!
Let’s look at the Dihedral Group $D_4$ (symmetries of a square).
In $D_4 = \langle r,s \mid r^4=e,\; s^2=e,\; sr=r^{-1}s\rangle$, $H=\langle s\rangle=\{e,s\}, \qquad K=\langle rs\rangle=\{e,rs\}.$
Then, $HK=\{hk\mid h\in H,\ k\in K\}.$
Computing all products: $\begin{aligned} e\cdot e &= e,\\ e\cdot rs &= rs,\\ s\cdot e &= s,\\ s\cdot rs &= (sr)s. \end{aligned}$
Using $sr=r^{-1}s = r^3s$, we get $(sr)s=(r^3s)s=r^3.$
Hence, $HK = \{e,s,rs,r^3\}.$ Is $HK$ a subgroup? To be a subgroup it must be closed under multiplication.
Consider $r^3\cdot s \notin HK = \{e,s,rs,r^3\}$, even though both $r^3$ and $s$ belong to $HK.$
Theorem: Let $H$ and $K$ be subgroups of $G$. The product $HK$ is a subgroup of $G$ if and only if $HK = KH$.
Important remark: $HK = KH$ does NOT mean that every element commutes ($hk = kh$). It simply means that for any $h_1 k_1 \in HK$, there exists some possibly different combination $k_2 h_2 \in KH$ that yields the exact same final result. The sets are equal, not necessarily the elements.
Proof:
This requires a double-implication proof. We will use a helpful lemma: For any subgroup $S \le G$, the set of all its inverses is exactly the subgroup itself ($S^{-1} = S$).
Part 1 (Forward): If $HK$ is a subgroup, then $HK = KH$.
Assume $HK \le G$. By our lemma, $(HK)^{-1} = HK$.
By the Shoes and Socks principle, the inverse of a product set reverses the order: $(HK)^{-1} = K^{-1}H^{-1}$
Because $H$ and $K$ are individual subgroups, $K^{-1} = K$ and $H^{-1} = H$.
Substituting this in, we get: $K^{-1}H^{-1} = KH$ Therefore, $HK = KH$.
Part 2 (Reverse): If $HK = KH$, then $HK$ is a subgroup.
Assume the sets commute as a whole ($HK = KH$). We will use the One-Step Subgroup test to prove $HK \le G$.
Let $x, y \in HK$. This means $x = h_1 k_1$ and $y = h_2 k_2$.
We must prove $xy^{-1} \in HK$. $x y^{-1} = (h_1 k_1)(h_2 k_2)^{-1} = h_1 k_1 k_2^{-1} h_2^{-1}$
Because $K$ is a subgroup, $k_1 k_2^{-1}$ is just some new element $k_3 \in K$. Because $H$ is a subgroup, $h_2^{-1}$ is just some new element $h_3 \in H$. $x y^{-1} = h_1 (k_3 h_3)$
Here is the magic step! The chunk $(k_3 h_3)$ belongs to the set $KH$. Because we assumed $KH = HK$, there must exist some combination $h_4 k_4 \in HK$ that is exactly equal to $k_3 h_3$.
We substitute it in: $x y^{-1} = h_1 (h_4 k_4) = (h_1 h_4) k_4$
Because $H$ is a subgroup, $h_1 h_4$ is just a single element in $H$. Our final result is exactly in the form (element of $H$) $\times$ (element of $K$).
Therefore, $xy^{-1} \in HK$. By the One-Step test, $HK$ is a valid subgroup. $\blacksquare$
Proposition: If $H$ is a subgroup of $G$, then $H^{-1} = H$.
Note: $H^{-1}$ is defined as the set of all inverses of elements in $H$: $\{h^{-1} \mid h \in H\}$.
Proof:
We prove this by double inclusion.
- Firstly, we are going to prove $H \subseteq H^{-1}$: Let $h \in H$. Because $H$ is a subgroup, it contains the inverse of every element, so $h^{-1} \in H$. The inverse of this inverse is $(h^{-1})^{-1} = h$. Because $h$ is the inverse of an element in $H$, it belongs to the set $H^{-1}$.
- Secondly, we need to prove $H^{-1} \subseteq H$: Let $x \in H^{-1}$. By definition, $x = h^{-1}$ for some $h \in H$. Because $H$ is a subgroup, it is closed under inversion, so $h^{-1}$ must also live inside $H$. Thus, $x \in H$. $\blacksquare$
Warning: The converse is not true! Just because a subset equals its inverse set does not mean it is a subgroup. Let $G = \{1, -1, i, -i\}$ under complex multiplication. Let $S = \{i, -i\}$. The inverse set is $S^{-1} = \{i^{-1}, (-i)^{-1}\} = \{-i, i\} = S$. However, $S$ is not a subgroup! It lacks the identity element ($1 \notin S$), and it fails closure ($i \cdot (-i) = 1 \notin S$).
Bibliography
- NPTEL-NOC IITM, Introduction to Galois Theory.
- Algebra, Second Edition, by Michael Artin.
- LibreTexts, Abstract and Geometric Algebra, Abstract Algebra: Theory and Applications (Judson).
- Field and Galois Theory, by Patrick Morandi. Springer.
- Michael Penn (Abstract Algebra), and MathMajor.
- Contemporary Abstract Algebra, Joseph, A. Gallian.
- Andrew Misseldine: College Algebra and Abstract Algebra.

