Skip to content

Find the exact object.

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

Exercise: reconstruct a finite tensor product

This problem turns a character table into an invariant-coupling calculation. Work over C\mathbb C in characteristic zero, with ordinary linear representations. A singlet below means the trivial representation, not an arbitrary one-dimensional irrep. No matrix basis is assumed until an explicit polynomial is requested.

Read characters and reality and finite tensor products first. The finite-symmetry path places this exercise between the mathematical method and its physical interpretation.

Let W={(x1,x2,x3)∈C3:x1+x2+x3=0}W=\{(x_1,x_2,x_3)\in\mathbb C^3:x_1+x_2+x_3=0\}, with S3S_3 permuting the coordinates. Use the following table, obtained by subtracting the invariant line from the permutation action. Its order and row labels match the S3 atlas record.

Class 11 (12)(12) (123)(123)
Size 11 33 22
Square class 11 11 (123)(123)
χ1\chi_{\mathbf1} 11 11 11
χsgn\chi_{\mathrm{sgn}} 11 −1-1 11
χW\chi_W 22 00 −1-1

The last square-class entry means the class of all three-cycles: (123)2=(132)(123)^2=(132) is a different element in the same class. The table agrees with Etingof et al., §4.8, printed p. 74.

  1. Verify irreducibility of WW and completeness of the three rows.
  2. Decompose W⊗WW\otimes W, Sym⁡2W\operatorname{Sym}^2W, and Λ2W\Lambda^2W. Check every character vector and total dimension.
  3. Compute the Frobenius–Schur indicator of WW.
  4. How many quadratic invariant polynomials does one commuting scalar multiplet in WW have? Write one on the plane x1+x2+x3=0x_1+x_2+x_3=0.
  5. Optional: find the invariant multiplicity in Sym⁡3W\operatorname{Sym}^3W and write a cubic invariant.

Use the normalized class inner product

⟨χ,ψ⟩=16∑C∣C∣χ(C)ψ(C)‾.\langle\chi,\psi\rangle =\frac16\sum_C |C|\chi(C)\overline{\psi(C)}.

For the square, exchange of tensor factors separates equal and unequal basis indices. If the eigenvalues of gg on VV are ziz_i, then ∑i,jzizj=(∑izi)2\sum_{i,j}z_iz_j=(\sum_i z_i)^2 and the diagonal contribution is ∑izi2=χV(g2)\sum_i z_i^2=\chi_V(g^2). Hence

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.

For the optional cubic, average the S3S_3 permutations of three tensor factors: there are one identity, three transpositions, and two three-cycles. Their traces are respectively χ(g)3\chi(g)^3, χ(g)χ(g2)\chi(g)\chi(g^2), and χ(g3)\chi(g^3). Thus

χSym⁡3V(g)=χ(g)3+3χ(g)χ(g2)+2χ(g3)6.\chi_{\operatorname{Sym}^3V}(g) =\frac{\chi(g)^3+3\chi(g)\chi(g^2)+2\chi(g^3)}6.

The character inner-product method is in Etingof et al., §§4.5 and 4.9; GAP’s symmetrization documentation, §72.11 explains why character values and power maps are separate inputs.

The norm of the standard character is (4+0+2)/6=1(4+0+2)/6=1. Its inner products with the trivial and sign rows are both (2−2)/6=0(2-2)/6=0. The two linear rows have norm one and inner product (1−3+2)/6=0(1-3+2)/6=0. Thus the character Gram matrix is the identity. The three constructed irreps satisfy 12+12+22=6=∣S3∣1^2+1^2+2^2=6=|S_3|, so no further irrep can appear in the regular representation. This uses complete reducibility in the declared category.

Pointwise multiplication gives χW⊗W=(4,0,1)\chi_{W\otimes W}=(4,0,1). Its inner products with the rows are

m1=(4+0+2)/6=1,msgn=(4+0+2)/6=1,mW=(8+0−2)/6=1.\begin{aligned} m_{\mathbf1}&=(4+0+2)/6=1,\\ m_{\mathrm{sgn}}&=(4+0+2)/6=1,\\ m_W&=(8+0-2)/6=1. \end{aligned}

Therefore W⊗W=1⊕sgn⊕WW\otimes W=\mathbf1\oplus\mathrm{sgn}\oplus W. Adding the rows gives (1,1,1)+(1,−1,1)+(2,0,−1)=(4,0,1)(1,1,1)+(1,-1,1)+(2,0,-1)=(4,0,1), which reconstructs every character value. The dimension is 4=1+1+24=1+1+2.

The square map supplies χW(g2)=(2,2,−1)\chi_W(g^2)=(2,2,-1), so

χSym⁡2W=(3,1,0),χΛ2W=(1,−1,1).\chi_{\operatorname{Sym}^2W}=(3,1,0),\qquad \chi_{\Lambda^2W}=(1,-1,1).

For the symmetric square, the three multiplicities are respectively (3+3)/6=1(3+3)/6=1, (3−3)/6=0(3-3)/6=0, and 6/6=16/6=1. The exterior character is exactly the sign row. Consequently

Sym⁡2W=1⊕W,Λ2W=sgn.\boxed{\operatorname{Sym}^2W=\mathbf1\oplus W,\qquad \Lambda^2W=\mathrm{sgn}.}

The dimensions are 3=2(2+1)/23=2(2+1)/2 and 1=2(2−1)/21=2(2-1)/2, adding to the tensor square’s dimension 44. Independently, Λ2W\Lambda^2W is the determinant of the plane action. The full permutation representation has determinant sgn\mathrm{sgn} and its fixed line has determinant one, confirming the result.

The Frobenius–Schur indicator is

ν(W)=1∣S3∣∑g∈S3χW(g2)=2+3⋅2+2(−1)6=1.\nu(W)=\frac1{|S_3|}\sum_{g\in S_3}\chi_W(g^2) =\frac{2+3\cdot2+2(-1)}6=1.

Thus WW has real type, consistently with its realization as the complexification of a real plane. The indicator convention is that of Etingof et al., §5.1, especially the calculation on printed p. 95.

A homogeneous quadratic polynomial lies in Sym⁡2(W∗)\operatorname{Sym}^2(W^*). The invariant dimension equals that in Sym⁡2W\operatorname{Sym}^2W; here W∗≃WW^*\simeq W as well. It is one, represented by

q=x12+x22+x32=2(x12+x1x2+x22).q=x_1^2+x_2^2+x_3^2=2(x_1^2+x_1x_2+x_2^2).

On the real plane this is a positive quadratic form. On the complexified field space it is a holomorphic polynomial, distinct from the Hermitian form ∑i∣xi∣2\sum_i|x_i|^2. Symmetry does not fix its coefficient in a potential. The exterior-square sign channel is neither a singlet nor an extra quadratic scalar coupling of one commuting multiplet.

The cube map gives χW(g3)=(2,0,2)\chi_W(g^3)=(2,0,2). Substitution into the cubic formula gives (4,0,1)(4,0,1), so

Sym⁡3W=1⊕sgn⊕W.\operatorname{Sym}^3W=\mathbf1\oplus\mathrm{sgn}\oplus W.

Its dimension is 4=(2+3−13)4=\binom{2+3-1}{3}. The unique trivial channel has the explicit representative

c=x13+x23+x33=3x1x2x3,x1+x2+x3=0.c=x_1^3+x_2^3+x_3^3=3x_1x_2x_3, \qquad x_1+x_2+x_3=0.

The equality follows by expanding (x1+x2+x3)3(x_1+x_2+x_3)^3. Both expressions are symmetric under coordinate permutations, verifying the invariant directly.

Problem B: why the Q₈ doublet behaves differently

Section titled “Problem B: why the Q₈ doublet behaves differently”

Write Q8={1,z,±u,±v,±w}Q_8=\{1,z,\pm u,\pm v,\pm w\}, with z=−1z=-1, u2=v2=w2=zu^2=v^2=w^2=z, and uv=w=−vuuv=w=-vu. Let DD be the faithful complex doublet. The class order and linear-character labels below are those of the Q8 record, with u,v,wu,v,w its quaternionic generators.

Class 11 zz {±u}\{\pm u\} {±v}\{\pm v\} {±w}\{\pm w\}
Size 11 11 22 22 22
Square class 11 11 zz zz zz
1\mathbf1 11 11 11 11 11
aa 11 11 11 −1-1 −1-1
bb 11 11 −1-1 11 −1-1
cc 11 11 −1-1 −1-1 11
DD 22 −2-2 00 00 00

The linear characters factor through Q8/{1,z}≃C2×C2Q_8/\{1,z\}\simeq C_2\times C_2. The doublet can be realized by determinant-one matrices

ρ(u)=(i00−i),ρ(v)=(01−10),ρ(w)=ρ(u)ρ(v).\rho(u)=\begin{pmatrix}i&0\\0&-i\end{pmatrix},\qquad \rho(v)=\begin{pmatrix}0&1\\-1&0\end{pmatrix},\qquad \rho(w)=\rho(u)\rho(v).

Squaring the first two matrices and multiplying them in both orders verifies the presentation; taking traces produces the last row. This construction and table are treated in Etingof et al., §§4.3(3) and 4.8.

Compute D⊗DD\otimes D, Sym⁡2D\operatorname{Sym}^2D, Λ2D\Lambda^2D, and ν(D)\nu(D). Does one commuting complex scalar in DD have a holomorphic quadratic invariant? Does it have a Hermitian quadratic invariant?

The doublet norm is (4+4)/8=1(4+4)/8=1, and the squared dimensions sum to 1+1+1+1+4=81+1+1+1+4=8. The tensor-square character is (4,4,0,0,0)(4,4,0,0,0). Its multiplicity against each linear row is (4+4)/8=1(4+4)/8=1, while its multiplicity against DD is (8−8)/8=0(8-8)/8=0. Therefore

D⊗D=1⊕a⊕b⊕c.D\otimes D=\mathbf1\oplus a\oplus b\oplus c.

The square map gives χD(g2)=(2,2,−2,−2,−2)\chi_D(g^2)=(2,2,-2,-2,-2). Hence

χSym⁡2D=(3,3,−1,−1,−1)=χa+χb+χc,χΛ2D=(1,1,1,1,1)=χ1.\begin{aligned} \chi_{\operatorname{Sym}^2D}&=(3,3,-1,-1,-1)=\chi_a+\chi_b+\chi_c,\\ \chi_{\Lambda^2D}&=(1,1,1,1,1)=\chi_{\mathbf1}. \end{aligned}

This checks all columns, not only the dimensions. The symmetric and exterior dimensions are respectively 33 and 11, and

ν(D)=2+2+2(−2)+2(−2)+2(−2)8=−1.\nu(D)=\frac{2+2+2(-2)+2(-2)+2(-2)}8=-1.

So the doublet is quaternionic type despite its real-valued character. Its singlet lies in the antisymmetric square. This also follows from the matrices: their determinants are one, so they preserve the alternating form ϵ=(01−10)\epsilon=\begin{pmatrix}0&1\\-1&0\end{pmatrix}, which spans Λ2D∗\Lambda^2D^*.

For commuting components, ϕTϵϕ=ϕ1ϕ2−ϕ2ϕ1=0\phi^T\epsilon\phi=\phi_1\phi_2-\phi_2\phi_1=0. There is no holomorphic quadratic invariant of a single doublet. By contrast, ϕ†ϕ\phi^\dagger\phi uses D∗⊗DD^*\otimes D and is invariant under these unitary matrices. Distinct scalar species can also have the nonvanishing holomorphic contraction ϕTϵη\phi^T\epsilon\eta. No statement about a relativistic fermion bilinear follows from the internal exterior square alone; its Lorentz and species indices must also be included.

The solutions above were derived from the tables and group relations. Use the workbench to check the implementation and retain its exact exports:

Calculation Expected result
S3 tensor square 1⊕sgn⊕W\mathbf1\oplus\mathrm{sgn}\oplus W
S3 symmetric square 1⊕W\mathbf1\oplus W
S3 exterior square sgn\mathrm{sgn}
S3 symmetric cube 1⊕sgn⊕W\mathbf1\oplus\mathrm{sgn}\oplus W
Q8 symmetric square a⊕b⊕ca\oplus b\oplus c
Q8 exterior square 1\mathbf1

For the method in reusable notation, continue to finite-character formulas. For an application with a complete field specification, continue to finite internal symmetries.

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