Skip to content

Find the exact object.

Names, aliases, Unicode, and common TeX notation. Open full search

Finite characters: conventions and exact methods

The finite-character workbench decomposes representations of C3C_3, S3S_3, D8D_8, Q8Q_8, A4A_4, S4S_4, and A5A_5. Its category is finite-dimensional ordinary complex representations in characteristic zero. Every irreducible character is included for each of these seven groups. The declared scope concerns these groups and operations; the broader Atlas contains many additional reference records.

For a representation VV with character χV\chi_V, the multiplicity of an irrep RR is

mR(V)=1∣G∣∑C∣C∣ χV(C) χR(C)‾.m_R(V)=\frac{1}{|G|}\sum_C |C|\,\chi_V(C)\,\overline{\chi_R(C)}.

The sum runs over conjugacy classes. Tensor products multiply their characters, so

mR(V⊗W)=1∣G∣∑C∣C∣ χV(C)χW(C)χR(C)‾.m_R(V\otimes W) =\frac{1}{|G|}\sum_C |C|\,\chi_V(C)\chi_W(C)\overline{\chi_R(C)}.

These are the character orthogonality and multiplicity formulas discussed in Etingof et al., §§4.4–4.5 and 4.9 and implemented for ordinary characters by GAP’s scalar-product operations, §72.8. The website evaluates them locally using exact algebraic arithmetic.

Group Order Irreducible labels in this workbench Character values
C3C_3 3 1,χ,χ2\mathbf1,\chi,\chi^2 Z[ω]\mathbb Z[\omega]
S3S_3 6 1,sgn,2\mathbf1,\mathrm{sgn},\mathbf2 Integers
D8D_8 8 1,a,b,c,2\mathbf1,a,b,c,\mathbf2 Integers
Q8Q_8 8 1,a,b,c,2\mathbf1,a,b,c,\mathbf2 Integers
A4A_4 12 1,1′,1′′,3\mathbf1,\mathbf1',\mathbf1'',\mathbf3 Z[ω]\mathbb Z[\omega]
S4S_4 24 1,sgn,2,3,3′\mathbf1,\mathrm{sgn},\mathbf2,\mathbf3,\mathbf3' Integers
A5A_5 60 1,3,3′,4,5\mathbf1,\mathbf3,\mathbf3',\mathbf4,\mathbf5 Z[φ]\mathbb Z[\varphi]

Here

ω=e2πi/3,ω2+ω+1=0,φ=1+52,φ2−φ−1=0.\omega=e^{2\pi i/3},\qquad \omega^2+\omega+1=0, \qquad \varphi=\frac{1+\sqrt5}{2},\qquad\varphi^2-\varphi-1=0.

The character tables can be checked against the GroupNames character-table collection; each workbench table links to its individual source record. Conjugacy-class representatives are displayed because row and column labels can vary across references.

Three conventions deserve particular attention:

  • D8D_8 has order eight. It is the square-symmetry group with r4=s2=1r^4=s^2=1 and srs=r−1srs=r^{-1}. GroupNames denotes this group by D4D_4.
  • The two S4S_4 triplets are inequivalent. Here 3\mathbf3 is the four-letter permutation representation with its invariant line removed, and 3′=sgn⊗3\mathbf3'=\mathrm{sgn}\otimes\mathbf3 is the orientation-preserving cubic-rotation representation. This convention fixes which triplet a dimension label denotes. See Etingof et al., §4.8.
  • The two A5A_5 triplets are separately of real type. The prime does not mean complex conjugation. With c=(12345)c=(12345), our 3\mathbf3 takes character value φ\varphi on cc, while 3′\mathbf3' takes 1−φ1-\varphi. The two rows are displayed in GroupNames: A5A_5.

The interface selects an irrep and allows 1–12 copies of it as a direct sum. It tensors that sum with a second selected irrep, or takes its tensor, symmetric, or exterior power with degree 0≤n≤120\leq n\leq12; the input dimension must be at most 60. The underlying exact engine additionally accepts mixed finite direct sums, with each input multiplicity at most 12 and each input dimension at most 60. These bounds control released calculations; they do not bound the underlying mathematical representation theory.

A power map records the class containing gkg^k for each conjugacy class of gg. This information is needed for symmetric and exterior powers. It cannot in general be recovered uniquely from the matrix of irreducible character values. The GAP manual, §73.1, makes this distinction explicit.

The pair D8,Q8D_8,Q_8 provides an immediate example. With the columns ordered as

D81r2rsrsQ81−1ijk\begin{array}{c|ccccc} D_8&1&r^2&r&s&rs\\ Q_8&1&-1&i&j&k \end{array}

the two character matrices agree. Their squaring maps differ: reflections in D8D_8 square to the identity, whereas i,j,ki,j,k in Q8Q_8 square to −1-1. Consequently, identical tensor-product rules coexist with different symmetric squares, exterior squares, and reality types. The group presentations and table data appear in GroupNames: D4D_4 and GroupNames: Q8Q_8.

The current dataset stores a complete power cycle for each class representative. Powers from 1 through 12 are exposed with the table; periodicity determines every positive power.

If the eigenvalues of ρV(g)\rho_V(g) are α1,…,αd\alpha_1,\ldots,\alpha_d, define the formal generating functions

Hg(t)=∏j=1d11−tαj,Eg(t)=∏j=1d(1+tαj).H_g(t)=\prod_{j=1}^d\frac{1}{1-t\alpha_j}, \qquad E_g(t)=\prod_{j=1}^d(1+t\alpha_j).

Their coefficients are the characters of Sym⁡nV\operatorname{Sym}^n V and ΛnV\Lambda^nV. Logarithmic differentiation gives

nhn(g)=∑k=1nχV(gk)hn−k(g),nen(g)=∑k=1n(−1)k−1χV(gk)en−k(g),h0=e0=1.\begin{aligned} n h_n(g)&=\sum_{k=1}^n\chi_V(g^k)h_{n-k}(g),\\ n e_n(g)&=\sum_{k=1}^n(-1)^{k-1}\chi_V(g^k)e_{n-k}(g), \end{aligned} \qquad h_0=e_0=1.

These recurrences give the class characters directly. A subsequent character inner product gives the full irrep decomposition. At degree two they reduce to

χSym⁡2V(g)=χV(g)2+χV(g2)2,χΛ2V(g)=χV(g)2−χV(g2)2.\chi_{\operatorname{Sym}^2V}(g) =\frac{\chi_V(g)^2+\chi_V(g^2)}{2}, \qquad \chi_{\Lambda^2V}(g) =\frac{\chi_V(g)^2-\chi_V(g^2)}{2}.

The result is checked against

dim⁡(V⊗n)=dn,dim⁡Sym⁡nV=(d+n−1n),dim⁡ΛnV=(dn).\dim(V^{\otimes n})=d^n,\qquad \dim\operatorname{Sym}^nV=\binom{d+n-1}{n},\qquad \dim\Lambda^nV=\binom dn.

The zeroth power is the trivial representation. An exterior power with n>dn>d is the zero representation. These conventions are handled explicitly. The equivalent symmetrization operations and independent A5A_5 cubic examples are documented in GAP, §72.11.

Reality and the Frobenius–Schur indicator

Section titled “Reality and the Frobenius–Schur indicator”

For an irreducible ordinary complex representation, the second indicator is

ν(R)=1∣G∣∑C∣C∣ χR(C2)={+1,R has real type,0,R has complex type,−1,R has quaternionic type.\nu(R)=\frac{1}{|G|}\sum_C |C|\,\chi_R(C^2) =\begin{cases} +1,&R\text{ has real type},\\ 0,&R\text{ has complex type},\\ -1,&R\text{ has quaternionic type}. \end{cases}

An irrep of real type admits a nondegenerate invariant symmetric complex bilinear form. An irrep of quaternionic type admits an alternating one. Both have real-valued characters. The indicator therefore contains information beyond whether displayed character values are real. See Etingof et al., §5.1 and GAP, §71.12-5.

For an irrep, this can also be read as

ν(R)=dim⁡(Sym⁡2R)G−dim⁡(Λ2R)G.\nu(R)=\dim(\operatorname{Sym}^2 R)^G-\dim(\Lambda^2R)^G.

For a reducible representation the same character average is additive, but its value does not supply a single irrep reality classification. The table’s reality column always refers to irreducible constituents.

The algebraic values are stored as integer coefficient pairs a+bωa+b\omega or a+bφa+b\varphi. Products use the quadratic relations above. In the cubic-root field, complex conjugation acts as

a+bω‾=(a−b)−bω.\overline{a+b\omega}=(a-b)-b\omega.

Integer coefficients use arbitrary-precision arithmetic. Every division in the power recurrence and multiplicity formula is checked to be exact. Multiplicities are then required to be nonnegative integers. This avoids selecting an integer by rounding an approximate algebraic sum.

Validation checks class sizes, the sum of squared dimensions, row and column orthogonality, conjugate pairs, and Frobenius–Schur indicators. Independent finite-group enumeration checks every released class power map. Each calculated decomposition must reproduce the original character on every class and conserve total dimension. The test fixtures also include the A5A_5 symmetric/exterior cubes printed in the GAP manual and the contrasting D8,Q8D_8,Q_8 squares.

The workbench records the convention profile finite-ordinary-complex-v1, version 1.0.0, the character dataset version, algorithm version, inputs, and source locators. Retain the exported result when using a calculation in research; an ordinary share link uses the currently published engine. See reproducibility and citation.

The output gives multiplicities and invariant counts. Explicit contraction tensors require a chosen basis and representation model. Modular, projective, and real representation categories require their own data and operations. Independent human expert review remains recorded separately from these source and algorithm checks.

Continue with From characters to physical couplings for worked examples, or open the finite-character workbench.

First public edition · mathematical baseline 6.0.0. Source curation and automated validation are distinct from independent expert review.

About · Contributors · Review status · Suggest a correction · Data and reproducibility · Articles CC BY 4.0 · Code MIT