simple group



Standard definition

A simple group is a group GG with exactly two quotient groups: the trivial quotient group {1}G/G\{1\} \cong G/G and the group GG/{1}G \cong G/\{1\} itself.

Equivalently, a simple group is a group possessing exactly two normal subgroups: the trivial subgroup {1}\{1\} and the group GG itself. One can also say that a normal subgroup is trivial iff it is not GG, or trivial iff proper (compare the definition in constructive mathematics below).

Note that the trivial group does not itself count as simple, on the grounds that it has only one quotient group (or only one normal subgroup). It may be possible to find authors that use “at most” in place of “exactly”, thereby allowing the trivial group to be simple. (Compare too simple to be simple.)

In constructive algebra

In constructive mathematics, we consider a group GG equipped with a tight apartness \ne such that the group operations are strongly extensional and use the theory of antisubgroups (which classically are the complements of subgroups). Then GG is simple if, given any normal antisubgroup NN of GG, NN is trivial iff it is proper. Explicitly, this says NN owns every nonidentity element (every xx such that x1x \ne 1) iff NN is inhabited (with some xx such that necessarily x1x \ne 1). Replacing ‘iff’ with ‘if’ here would allow the trival group to be simple.


Finite simple groups

Simple groups are most commonly encountered in the theory of finite groups. Every finite group GG admits a composition series, i.e., a finite filtration of subgroups

1=G 0G 1G n=G1 = G_0 \subseteq G_1 \subseteq \ldots \subseteq G_n = G

where each inclusion G iG i+1G_i \subseteq G_{i+1} is a normal subgroup and the quotient G i+1/G iG_{i+1}/G_i (called a composition factor) is simple. The condition of simplicity means that that the filtration cannot be further refined by addition of strict inclusions of normal subgroups. Furthermore, the Jordan-Hölder theorem ensures that any two composition series have the same length and the same composition factors (up to permutation).

Thus finite simple groups are in some sense the primitive building blocks of finite groups generally. The massive program of classifying all finite simple groups was announced as completed by Daniel Gorenstein in 1983, although some doubts remained because there were some gaps in proofs. Most if not all the gaps are considered by experts in the area to have been filled, but there remain some notable skeptics, including for example Jean-Pierre Serre, who said in an interview

Whenever I asked the specialists, they replied something like: “Oh no, it is not a gap; it is just something which has not been written, but there is an incomplete unpublished 800-page manuscript on it.” For me, it was just the same as a “gap,” and I could not understand why it was not acknowledged as such. (Source)

and possibly also John H. Conway, although according to Joe Shipman,

A few months ago I was discussing the COFSG with John Conway and he was still pessimistic. (Meaning of course that he was confident the classification was correct, “optimistic” in his case means there are previously undiscovered finite simple groups, because they’re beautiful and interesting objects and it would be disappointing to have no more.)

See classification of finite simple groups.

Directed colimits


Let S αS_\alpha be a directed system of simple groups and monomorphisms between them. Then colim αS αcolim_\alpha S_\alpha is also simple.


Suppose NN is a normal subgroup of colim αS αcolim_\alpha S_\alpha with a non-identity element. Then N α=NS αN_\alpha = N \cap S_\alpha is normal in S αS_\alpha (if xS αx \in S_\alpha and yNS αy \in N \cap S_\alpha, then clearly xyx 1x y x^{-1} belongs to both NN and S αS_\alpha). As soon as α\alpha is large enough that S αS_\alpha contains a non-identity element of NN, it follows from simplicity of S αS_\alpha that N α=S αN_\alpha = S_\alpha. By directedness, this shows N α=S αN_\alpha = S_\alpha for all α\alpha, and since for any sSs \in S there is α\alpha such that sS αs \in S_\alpha, we conclude sS α=N αNs \in S_\alpha = N_\alpha \subseteq N, i.e., NN contains any sSs \in S.

Infinite simple groups

There are simple groups of any infinite cardinality κ\kappa; take for example the smallest normal subgroup of the automorphism group Aut(κ)Aut(\kappa) containing all 3-cycles (this is the infinite version of the alternating group).

To see that Alt(κ)Alt(\kappa) is simple, it is enough to observe that it is a directed colimit of Alt(X)Alt(X) where XX ranges over finite subsets of κ\kappa of cardinality at least 55; then simplicity of A nA_n for integers n0n \geq 0, coupled with Proposition 1, yields the desired result.

A rather deeper set of examples is afforded by the following result, which follows from a theorem named after Baer, Ulam, and Schreier.


For an infinite set XX, every proper normal subgroup of the permutation group Sym(X)Sym(X) is contained in a maximal such normal subgroup N XN_X, consisting of all permutations that move fewer than |X|{|X|} many elements of XX, where |X|{|X|} is the cardinality of XX.

It follows that the quotient group Q X=Sym(X)/N XQ_X = Sym(X)/N_X is simple, and we have the following corollary.


Every group embeds into a simple group.


For finite groups GG (WLOG, of cardinality at least 33), we have the Cayley embedding GSym(G)G \hookrightarrow Sym(G) into the permutation group of the underlying set, and there is an embedding Sym(G)Alt(G+2)Sym(G) \hookrightarrow Alt(G + 2) which carries an even permutation on GG to the obvious even permutation on G+2G + 2 that fixes the elements a,ba, b of 22, and an odd permutation π\pi on GG to the even permutation π(ab)\pi (a\; b). Hence GG embeds in a simple group Alt(G+2)Alt(G + 2).

An infinite group GG embeds in Q GQ_G via the Cayley embedding

GSym(G)Sym(G)/N G=Q GG \hookrightarrow Sym(G) \to Sym(G)/N_G = Q_G

noting that for any non-identity gGg \in G, the permutation Cayley(g)=(hgh)Cayley(g) = (h \mapsto g h) has no fixed points, hence does not belong to N GN_G, so that GQ GG \to Q_G is indeed monic.

Tarski monsters

Among the infinite simple groups, there are curious examples called “Tarski monsters” for primes pp, infinite groups with the property that every subgroup is either trivial, of prime order pp, or the entire group. Their existence (for all primes p>10 75p \gt 10^{75}, according to Wikipedia) was established only in 1979, by Olshanskii. It may be shown that all such groups must be simple, and generated by two elements, and also that for each prime p>10 75p \gt 10^{75} there are continuum many monsters for pp that are distinct up to isomorphism.

Last revised on March 27, 2018 at 19:31:04. See the history of this page for a list of all contributions to it.