Matrix groups over finite fields — the 16 infinite families
The groups of Lie type are built from finite fields. A finite field (or Galois field) GF(q) exists for every prime power q = pⁿ, and these are the only finite fields up to isomorphism. The simplest examples are Z/pZ for p prime, but when n > 1 the construction requires adjoining roots of irreducible polynomials over Z/pZ. Every finite field has a cyclic multiplicative group of order q − 1, and their rich algebraic structure provides exactly the raw material needed to construct matrix groups.
The connection to simple groups arises because we can form matrix groups over finite fields — groups like GL(n, q), the group of n × n invertible matrices with entries in GF(q). These groups are large and highly structured, and by taking appropriate quotients and subgroups we extract simple groups. The process mirrors the classical theory of Lie groups (continuous symmetry groups) but replaces the real or complex numbers with a finite field, producing finite analogues of the great classical groups.
The general linear group GL(n, q) itself is not simple — its center, consisting of scalar matrices, is a non-trivial normal subgroup. The special linear groupSL(n, q) (matrices of determinant 1) removes one obstruction, but the scalar matrices of determinant 1 still form a central subgroup. Dividing out this center yields the projective special linear group PSL(n, q) = SL(n, q) / Z(SL(n, q)), and this group is simple in most cases — specifically, for all n ≥ 2 and q ≥ 2 except for PSL(2, 2) ≅ S₃ and PSL(2, 3) ≅ A₄.
The groups of Lie type fall into 16 infinite families, which can be organized into three categories. The classical groups come from the standard matrix groups: PSL(n, q) (type Aₙ₋₁), PSp(2n, q) (type Cₙ, symplectic groups), PSO(2n+1, q) (type Bₙ, odd-dimensional orthogonal), and PSO⁺(2n, q) and PSO⁻(2n, q) (type Dₙ, even-dimensional orthogonal). These generalize the classical Lie groups SL(n), Sp(2n), and SO(n) over the reals.
The exceptional groups correspond to the five exceptional Lie algebras: G₂(q), F₄(q), E₆(q), E₇(q), and E₈(q). These groups were constructed by Chevalley in his groundbreaking 1955 paper, which unified the theory by showing that every simple Lie algebra over the complex numbers gives rise to a family of simple groups over every finite field. Chevalley's construction was a watershed moment: it revealed that the "accidental" simple groups known to 19th-century mathematicians were instances of a systematic pattern.
The twisted groups arise from symmetries of Dynkin diagrams. When a Dynkin diagram has a non-trivial automorphism, we can "twist" the corresponding Chevalley group to produce a new family: the Steinberg groups ²Aₙ(q), ²Dₙ(q), ²E₆(q), and ³D₄(q), the Suzuki groups ²B₂(2²ⁿ⁺¹), and the Ree groups ²G₂(3²ⁿ⁺¹) and ²F₄(2²ⁿ⁺¹). Including the Tits group ²F₄(2)' (sometimes counted as sporadic), these twisted constructions complete the 16 families. Together with the cyclic and alternating groups, the Lie-type families account for all but 26 of the finite simple groups.
The interactive display below organizes the 16 families of Lie-type groups, showing their Dynkin diagram types, order formulas, and the relationships between classical, exceptional, and twisted families. Explore how the order of each group depends on the rank n and the field size q.
|PSL(2, q)| = q(q² − 1) / gcd(2, q − 1). These are simple for all prime powers q ≥ 4.
| q | |PSL(2,q)| | Isomorphic to |
|---|---|---|
| 4 | 60 | A₅ |
| 5 | 60 | A₅ |
| 7 | 168 | PSL(2,7) — 2nd smallest non-abelian simple |
| 8 | 504 | PSL(2,8) |
| 9 | 360 | A₆ |
| 11 | 660 | PSL(2,11) |
| 13 | 1,092 | PSL(2,13) |
| 16 | 4,080 | PSL(2,16) |
| 17 | 2,448 | PSL(2,17) |
| 19 | 3,420 | PSL(2,19) |
Key insight: The order formulas for groups of Lie type follow a beautiful pattern: |PSL(n, q)| = (1/d) · qⁿ⁽ⁿ⁻¹⁾/² · ∏ᵢ₌₂ⁿ (qⁱ − 1), where d = gcd(n, q − 1). As q grows, these groups become enormous. PSL(2, q) alone produces an infinite family of simple groups — one for each prime power q ≥ 4 — with orders q(q² − 1)/2.
The family PSL(2, q) is the most studied and best understood among the groups of Lie type. For q ≥ 4, PSL(2, q) is a simple group of order q(q − 1)(q + 1)/gcd(2, q − 1). These groups have a rich geometry: PSL(2, q) acts on the projective line over GF(q), which has q + 1 points, by Mobius transformations z ↦ (az + b)/(cz + d). This action is doubly transitive, making PSL(2, q) a key example in permutation group theory.
Several small groups of Lie type coincide with other known simple groups, revealing deep connections. PSL(2, 4) ≅ PSL(2, 5) ≅ A₅ (the icosahedral group, order 60). PSL(2, 7) ≅ PSL(3, 2) ≅ GL(3, 2) is the second smallest non-abelian simple group, with 168 elements — it is the automorphism group of the Fano plane, the simplest finite projective geometry. PSL(2, 9) ≅ A₆. These "exceptional isomorphisms" are low-dimensional accidents that do not generalize but illustrate the interconnectedness of the classification.
The subgroup structure and representation theory of PSL(2, q) are completely understood and serve as a model for the study of larger groups of Lie type. Dickson classified all subgroups of PSL(2, q) in 1901: they include cyclic groups, dihedral groups, A₄, S₄, A₅, and the Borel subgroup (upper triangular matrices). This complete understanding of PSL(2, q) was essential for the classification program, as many arguments reduce to the PSL(2, q) case by localizing at involutions and their centralizers.