Simple Groups

Simple groups are not a museum piece; they are load-bearing structure throughout modern mathematics and its applications. In Galois theory, the simplicity and non-abelian-ness of $A_5$ (and $A_n$, $n \geq 5$) is exactly what obstructs a general radical formula for degree-$5$ (and higher) polynomials — solvability by radicals corresponds to the Galois group being a solvable group, and $A_n$'s simplicity for $n \geq 5$ blocks this. In the classification of finite simple groups, itself, simple groups are literally the objects being enumerated: the classification names the cyclic groups of prime order, the alternating groups $A_n$ ($n\geq5$), 16 families of groups of Lie type, and 26 sporadic groups (including the Monster group), as the complete list. In cryptography, large simple groups (e.g. from families of Lie type over finite fields) underlie hard problems used in some group-based cryptographic schemes, and the rigidity of simple groups (lack of nontrivial quotients) is relevant to constructing pseudorandom generators and expander graphs from Cayley graphs of simple groups. In coding theory and combinatorics, sporadic simple groups such as the Mathieu groups arise as automorphism groups of highly symmetric combinatorial designs (Steiner systems), connecting simple groups to error-correcting codes. In physics, simple Lie groups (the continuous analogue) classify possible symmetry groups of gauge theories in the Standard Model, and the representation theory built on simple groups underlies particle classification schemes. Finally, in computational group theory, the Sylow-counting techniques of this article are directly implemented in computer algebra systems (GAP, Magma) to test group orders for possible simplicity as a first filtering step before deeper structural analysis.

Ask the tutor: How does the Jordan-Hölder theorem guarantee that composition factors are unique even though composition series themselves are not unique?

Think of the collection of all finite groups, partially organized by the relation "$N$ is a normal subgroup of $G$, so $G$ can be built from $N$ and $G/N$." A normal subgroup $N \neq \{e\}, G$ lets us break $G$ into two smaller pieces, $N$ and $G/N$, and in principle reconstruct information about $G$ from them (this reconstruction is the extension problem, and it is subtle — but the point is that $G$ is not "indecomposable"). A simple group is one where this breaking-down process cannot even start: there is no such intermediate normal subgroup, so the group is, in the relevant structural sense, indivisible. This is exactly parallel to prime factorization. A prime $p$ has no factorization $p = ab$ with $1 < a, b < p$; a simple group has no normal subgroup strictly between $\{e\}$ and $G$. Just as repeatedly factoring composite integers into primes terminates in a unique multiset of primes, repeatedly factoring a finite group along normal subgroups (a composition series) terminates in a unique multiset of simple groups (the composition factors), by the Jordan-Hölder theorem. The order in which you factor may differ, but the atoms you end up with do not. Where the analogy breaks down — and where the subject becomes vastly richer than number theory — is in reconstruction. Given the prime factors of $n$, you know $n$ completely: multiplication of integers is commutative and there's only one way to reassemble them. Given the composition factors of $G$, you generally do not know $G$: the same composition factors can be assembled into many non-isomorphic groups (this is the extension problem, governed by group cohomology). So simple groups are the alphabet, but the "words" built from them are far more varied than integers built from primes. The Sylow theorems become the primary elementary weapon for detecting when a group is not simple. Recall: if $|G| = p^a m$ with $p \nmid m$, the number $n_p$ of Sylow $p$-subgroups satisfies $n_p \equiv 1 \pmod p$ and $n_p \mid m$. If for some prime $p$ we can force $n_p = 1$, the unique Sylow $p$-subgroup is normal (and proper and nontrivial whenever $1 < p^a < |G|$), so $G$ is not simple. Even when $n_p > 1$ is forced, we can sometimes still find a normal subgroup by having $G$ act on the set of Sylow $p$-subgroups by conjugation, producing a homomorphism to a symmetric group whose kernel yields the required normal subgroup, or by counting elements directly.

Example 1: No simple group of order 56. Let $G$ be a group with $|G| = 56 = 2^3 \cdot 7$. We show $G$ is not simple by Sylow counting. By Sylow's theorems, the number $n_7$ of Sylow $7$-subgroups satisfies $n_7 \equiv 1 \pmod 7$ and $n_7 \mid 8$. The divisors of $8$ are $1,2,4,8$; among these, only $1$ and $8$ are $\equiv 1 \pmod 7$ (check: $8 = 7+1 \equiv 1$; $2 \equiv 2$; $4 \equiv 4$). So $n_7 \in \{1, 8\}$. Case $n_7 = 1$: the unique Sylow $7$-subgroup $P$ is normal (uniqueness plus Sylow's conjugacy theorem force every conjugate of $P$ to equal $P$), and $1 < |P| = 7 < 56$, so $P$ is a proper nontrivial normal subgroup. $G$ is not simple. Case $n_7 = 8$: Each Sylow $7$-subgroup has order $7$ (prime), so any two distinct Sylow $7$-subgroups intersect trivially (a nontrivial intersection would be a common subgroup of order dividing $7$, forcing the subgroups to coincide). Thus the $8$ Sylow $7$-subgroups contribute

distinct nonidentity elements (each contributes its $6$ nonidentity elements, and these sets are pairwise disjoint by the trivial-intersection property, since any nonidentity element lies in exactly one Sylow 7-subgroup — the one it generates). This accounts for $48$ elements of order $7$, plus the identity, totaling $49$ elements. The remaining

elements must contain all Sylow $2$-subgroups (each of order $8$), since a Sylow $2$-subgroup, having order $8$ coprime to $7$, can contain no element of order $7$, so it lies entirely among these remaining $7$ elements — but a subgroup of order $8$ needs $8$ elements, and only $7$ remain! Hence there can be at most $\lfloor 7/8 \rfloor = 0$... more precisely, there is room for at most one Sylow $2$-subgroup, because two distinct Sylow $2$-subgroups would each need $8$ elements from a pool of only $7$ available non-order-7 elements, an impossibility if they were disjoint outside identity, and even accounting for overlap at the identity, two distinct subgroups of order $8$ intersecting only in $\{e\}$ would need $8 + 8 - 1 = 15 > 7$ elements. Hence $n_2 = 1$. Therefore in Case $n_7=8$ we still get $n_2 = 1$, so the unique Sylow $2$-subgroup $Q$ (order $8$) is normal, proper, and nontrivial. In both cases $G$ has a proper nontrivial normal subgroup, so no group of order $56$ is simple.

Example 2: No simple group of order 36. Let $|G| = 36 = 2^2 \cdot 3^2$. We show $G$ is not simple, illustrating the "action on Sylow subgroups" technique for when Sylow counting alone does not immediately force $n_p = 1$. The number of Sylow $3$-subgroups satisfies $n_3 \equiv 1 \pmod 3$ and $n_3 \mid 4$. Divisors of $4$: $1,2,4$; those $\equiv 1 \pmod 3$: $1$ and $4$ (since $4 = 3+1$). So $n_3 \in \{1,4\}$. Case $n_3 = 1$: the Sylow $3$-subgroup is normal, proper ($|P|=9 < 36$), and nontrivial. Done. Case $n_3 = 4$: Let $\mathrm{Syl}_3(G) = \{P_1, P_2, P_3, P_4\}$ and let $G$ act on this $4$-element set by conjugation:

This action is transitive by Sylow's conjugacy theorem, in particular nontrivial, so $\varphi$ is not the trivial homomorphism. Let $K = \ker \varphi$. Since $\varphi$ is nontrivial, $K \neq G$. We must check $K \neq \{e\}$, i.e., $\varphi$ is not injective. If $\varphi$ were injective, $G$ would embed into $S_4$, but $|G| = 36 > 24 = |S_4|$ — impossible for an injective homomorphism (an injective map cannot increase then decrease cardinality; concretely $|G| = |\varphi(G)| \leq |S_4| = 24 < 36$, a contradiction). So $\varphi$ is not injective, giving $K \neq \{e\}$. Since $K = \ker \varphi \trianglelefteq G$ (kernels of homomorphisms are always normal) and $\{e\} \neq K \neq G$, $K$ is a proper nontrivial normal subgroup of $G$. In both cases, $G$ has a proper nontrivial normal subgroup, so no group of order $36$ is simple.