Group theory
The Sylow Theorems: Finding Prime-Power Subgroups
Lagrange's theorem tells you the order of a subgroup divides the order of the group — but it stays silent on which divisors are actually achieved. The Sylow theorems supply the sharpest possible partial converse: if a prime power pk divides |G|, then G contains a subgroup of order pk, and for the maximal such power the subgroups are all conjugate, tightly counted, and everywhere.
Precisely: let G be a finite group with |G| = pn·m where p is prime and p ∤ m. Then (I) G has a subgroup of order pn — a Sylow p-subgroup; (II) any two Sylow p-subgroups are conjugate and every p-subgroup sits inside one; and (III) the number np of Sylow p-subgroups satisfies np ≡ 1 (mod p) and np | m. These three facts turn abstract divisibility into concrete, countable structure.
- FieldFinite group theory
- Proved byLudwig Sylow, 1872
- Key hypothesisG finite; p prime with p^n ‖ |G|
- StatementSylow p-subgroups of order p^n exist, are all conjugate, and number n_p ≡ 1 (mod p), n_p | m
- Proof techniqueGroup actions on cosets/subsets + orbit-counting (class equation)
- GeneralizesCauchy's theorem (k=1) and the converse of Lagrange for prime powers
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
The precise statement
Fix a prime p and a finite group G, and write |G| = pn·m with p ∤ m. A subgroup of order pk for some k is a p-subgroup; a p-subgroup of the maximal order pn is a Sylow p-subgroup. Let Sylp(G) denote the set of all of them and np = |Sylp(G)|.
- Sylow I (Existence). Sylp(G) ≠ ∅. More strongly, for every k ≤ n there is a subgroup of order pk, and each is contained in one of order pk+1 (a full chain).
- Sylow II (Conjugacy). If P, Q ∈ Sylp(G) then Q = gPg−1 for some g ∈ G. Moreover every p-subgroup of G lies inside some Sylow p-subgroup.
- Sylow III (Counting). np ≡ 1 (mod p) and np divides m. Equivalently np = [G : NG(P)], the index of the normalizer of any Sylow p-subgroup.
The hypothesis that G is finite is essential; pn must be the exact power of p dividing |G|, not merely some power that divides it.
The picture: extracting the p-part
Think of |G| as factored into its prime pieces. Lagrange says subgroup orders divide |G|, but that's only a necessary condition — the alternating group A4 has order 12 yet no subgroup of order 6. Sylow rescues the situation for prime powers: the p-primary part of the group can always be realized as an honest subgroup, and up to relabeling by conjugation there is essentially one of them.
The right mental image is a group acting on a set and shuffling points into orbits. Orbit sizes divide |G|, so they carry p-adic information. Sylow's genius is to choose the acting set so cleverly that counting orbit sizes modulo p forces a large p-subgroup to materialize as the stabilizer of some point. Conjugacy (Sylow II) then says the acting group can slide any Sylow subgroup onto any other, so they form a single G-orbit; and the counting law (Sylow III) is just the orbit-stabilizer theorem applied to that orbit. Existence, uniqueness, and count are three readings of one action.
The key idea of the proof
Existence. Let Ω be the set of all subsets of G of size pn, and let G act by left translation. A Kummer/Lucas count shows |Ω| = C(pnm, pn) is not divisible by p (the p's cancel exactly). Since orbit sizes sum to |Ω|, some orbit has size prime to p. By orbit–stabilizer, the stabilizer H of a subset in that orbit has |G|/|orbit| divisible by pn, so pn | |H|; but H permutes a single pn-element set, forcing |H| ≤ pn. Hence |H| = pn — a Sylow subgroup.
Conjugacy and counting. Fix P ∈ Sylp(G) and let a p-subgroup Q act by left multiplication on the coset space G/P. Since |G/P| = m is coprime to p, and every Q-orbit has p-power size, there must be a fixed coset gP — which means Q ⊆ gPg−1. Taking Q a Sylow subgroup gives conjugacy. Letting P itself act by conjugation on Sylp(G), the only fixed point is P (a Sylow normalizer argument), so all other orbits have size divisible by p, giving np ≡ 1 (mod p). Orbit–stabilizer under the full G-action gives np = [G : NG(P)] | m.
Worked example: groups of order 15 and 12
Order 15 = 3·5. For p = 5: n5 ≡ 1 (mod 5) and n5 | 3, so n5 ∈ {1}. For p = 3: n3 ≡ 1 (mod 3) and n3 | 5, so n3 = 1. Both Sylow subgroups are unique, hence normal. A normal ℤ/5 and normal ℤ/3 with trivial intersection generate G ≅ ℤ/5 × ℤ/3 ≅ ℤ/15. Conclusion: every group of order 15 is cyclic — proved with no computation, purely from the counting law.
Order 12 = 2²·3. Here n3 ≡ 1 (mod 3) and n3 | 4, so n3 ∈ {1, 4}. If n3 = 4, the four Sylow 3-subgroups contribute 4·2 = 8 elements of order 3, leaving only 4 elements — exactly enough for a single (hence normal) Sylow 2-subgroup. So at least one of the Sylow subgroups is normal; a group of order 12 can never be simple. A4 realizes the n3 = 4 case, with its normal Klein four-group V4 as the Sylow 2-subgroup.
Why the hypotheses matter, and what breaks
Finiteness is indispensable. For infinite groups the statement is false as written: (ℚ, +) has no proper nontrivial finite p-subgroups at all, and the notion of a 'maximal p-power dividing |G|' is meaningless. Sylow theory lives in the finite world (with a locally-finite generalization requiring extra care).
Maximality of the power is the whole point. Sylow does not give a converse to Lagrange for all divisors. A4, of order 12, has no subgroup of order 6 even though 6 | 12 — because 6 is not a prime power. Sylow guarantees subgroups only of orders 1, 2, 4, 3 (the prime-power divisors), and indeed A4 has exactly those. Dropping 'prime power' collapses the theorem.
Relation to other results. Cauchy's theorem (1845) is precisely Sylow I with k = 1. The class equation and orbit–stabilizer are the engines. Sylow underpins the theory of p-groups (which are exactly their own Sylow subgroups), the Schur–Zassenhaus theorem, and the entire program of classifying finite simple groups.
Applications and significance
The Sylow theorems are the workhorse of finite group theory. Their most common use is a non-simplicity engine: to show a group of a given order must have a proper nontrivial normal subgroup, compute np; if the constraints np ≡ 1 (mod p) and np | m force np = 1, that Sylow subgroup is normal and the group is not simple. This single technique classifies groups of many small orders and rules out simple groups of countless orders — an indispensable step toward the classification of finite simple groups.
- Structure theorems: if all Sylow subgroups are normal, G is their direct product (nilpotent groups are exactly this case).
- Number-theoretic flavor: the counting congruence np ≡ 1 (mod p) is a group-theoretic cousin of results like Wilson's and Fermat's little theorem.
- Representation theory & modular forms lean on p-local structure that Sylow subgroups organize.
Whenever you know only |G| and want to deduce structure, Sylow is the first tool to reach for.
| Theorem | Claim | Mechanism | What it buys you |
|---|---|---|---|
| Sylow I (Existence) | A subgroup P ≤ G of order p^n exists | G acts on the set of p^n-element subsets; a p-divisibility count of binomial coefficients forces a fixed orbit | Guarantees maximal p-power subgroups always exist |
| Sylow II (Conjugacy / Domination) | All Sylow p-subgroups are conjugate; every p-subgroup lies in some Sylow p-subgroup | A p-group Q acts on the coset space G/P; a fixed point exists because |G/P| = m is prime to p | Makes 'the' Sylow p-subgroup canonical up to conjugacy; n_p = 1 ⇔ normal |
| Sylow III (Counting) | n_p ≡ 1 (mod p) and n_p divides m | P acts by conjugation on the set of Sylow p-subgroups; only P itself is fixed | Pins n_p to a tiny candidate list, often forcing normality |
| Cauchy (special case) | If p | |G| then G has an element of order p | Sylow I with k = 1, or a direct action of ℤ/p on p-tuples | Foundational existence result used inside Sylow's own proofs |
Frequently asked questions
Why must G be finite for the Sylow theorems?
The statement quantifies over the exact prime-power p^n dividing the finite number |G|, and the proofs count finite orbits modulo p. Infinite groups need not have a well-defined 'maximal p-part': (ℚ, +) is a torsion-free infinite group with no nontrivial finite p-subgroups at all. There are locally finite analogues, but they require additional hypotheses and are not the classical theorems.
Do the Sylow theorems give a converse to Lagrange's theorem?
Only a partial one, for prime-power divisors. Sylow guarantees a subgroup of order p^k for every prime power p^k dividing |G|, but says nothing about divisors that are not prime powers. The alternating group A₄ (order 12) has no subgroup of order 6, so the full converse of Lagrange is false — a fact Sylow does not contradict.
What does n_p ≡ 1 (mod p) actually let you prove?
Combined with n_p | m, the congruence pins the number of Sylow p-subgroups to a short list of candidates. Very often the only value satisfying both constraints is n_p = 1, which means the Sylow p-subgroup is unique and therefore normal. This is the standard route to proving a group of a given order cannot be simple.
How is Cauchy's theorem related to Sylow?
Cauchy's theorem (1845) states that if a prime p divides |G|, then G has an element — hence a subgroup — of order p. This is exactly Sylow I in the case k = 1. Historically Cauchy came first and is often used as a lemma inside proofs of Sylow's theorems, though Sylow can also be proved self-containedly via the subset-action argument.
Are all Sylow p-subgroups really isomorphic, or just conjugate?
Conjugate, which is stronger: if Q = gPg⁻¹ then conjugation by g is an isomorphism P → Q, so conjugate subgroups are automatically isomorphic. The converse fails in general — two isomorphic subgroups need not be conjugate — but for Sylow p-subgroups the theorem promises genuine conjugacy, so 'the' Sylow p-subgroup is well defined up to an inner automorphism of G.
When is a Sylow p-subgroup normal?
Exactly when n_p = 1. Since all Sylow p-subgroups are conjugate (Sylow II), there is a single one if and only if it is fixed by every conjugation, i.e. normal. So computing n_p and finding it forced to equal 1 is precisely a proof of normality; this is why Sylow III is the primary tool for detecting normal subgroups from order alone.