The Racah Algebra of Rank 2: Properties, Symmetries and Representation

We need to check only the non-central terms C_{12}, C_{23}, C_{13} and D_{123}. The first case to check is

Here, we have used (?) and the symmetry of D_{123} (?).

The second (and last) case to check is a pair of C’s with the D. Up to permutation, it is as follows:

We have shown in the rank-1 Racah algebra, the commutation relations satisfy the Jacobi identity without further constraints.

This follows directly from the commutation (2.3) and decomposition relations (2.3).

From (2.3) with K\rightarrow I \rightarrow J, we have

The resulting identity follows from the relations in Lemma 2.4.

This proof is directly analogous to Lemma 2.2 with indices replaced by subset. Relation (2.3) is a straightforward computation that follows from symmetry of the commutator

Relation (2.5) is less obvious. We check

Thus the triple commutation of any of the generators satisfies the Jacobi identity identically, as expected in an associative algebra.

Note that C_{k\ell} and hence P_{k\ell} will commute with both C_{ik\ell} and C_{jk\ell}. We use the decomposition relation (2.3) to obtain

Next we use the fact that D_{ijk} is invariant under cyclic permutations and odd under singleton flips D_{ijk}=-D_{ikj}. This leaves [C_{ik\ell}, C_{jk\ell}]=-2D_{ijk} -2D_{ij\ell}. So, finally we see that Lemma 2.7 holds from the commutativity of C_{k\ell} with C_{ik\ell} and C_{jk\ell}.

As described above, we used a symbolic manipulator (Maple [Maple]) to solve the linear system of equations involving terms P_{A} D_B for the required expressions. The resulting identities are non-trivial but can be checked as follows. On the one hand, it can be directly verified using the commutation (2.3) and decomposition relations (2.3) that [C_{ij}, D_{ijk}] can be expressed as

On the other hand, the first 2 lines are equal to 0 as consequences of the commutation relations (2.3), as C_{ijkl} and C_{ijk} will both commute with D_{ijk} (and permutations). The third line also vanishes from the relation in Lemma 2.7. Finally, the second to last line is comprised of terms containing “inner” quadratic relations for singleton subsets (?) and the last line “inner” quadratic relations for multi-index sets I=\{i\}, J=\{j\}, K=\{k\ell\} and I=\{i\}, J=\{k\}, K=\{j\ell\} respectively in (2.3). Replacing these last two lines with their quadratic identities leaves the resulting equation

which can be more succinctly expressed using the P generators as in (2.8).

Each of these are direct consequences of Lemma 2.8. We also note that the first equation is exactly Lemma 2.7 expressed in terms of P.

We use Lemma 2.4 to write [D_{ijk}, D_{jk\ell}] =\frac12 [D_{ijk}, [ P_{k\ell}, P_{\ell j}]] using the fact that our algebra is associative and so the Jacobi identity holds gives

We can remove the negative signs using the skew symmetry of the D’s to write the expression as

From the definition of the P_{ij} and P_{i} (?) and using the algebra decomposition relations (2.3), it is clear that the P’s generate the algebra and are symmetric under interchange of indices. Lemma 2.4 shows that the D_{ijk} are antisymmetric under interchange of indices.

Commutation relation (2.11) is the definition of the D’s while relation (2.11) follows directly from the change of basis applied to (3.1). Lemma 2.8 gives relation (2.11) and Lemma 2.10 gives (2.11). Finally, the identity (2.11) follows from the discussion above under a cyclic permutation, or equivalently using the Jacobi identity with [D_{jk\ell}, [P_{ik}, P_{k\ell}]].

It remains to consider Jacobi identities more generally. First, we begin with the case of three P’s. The Jacobi identity for triples \{P_{jk},P_{jk},P_{ij}\} and \{ P_{ki}, P_{ij}, P_{jk}\} are verified using Lemma 2.6. The case \{P_{jk},P_{jk},P_{i\ell}\} is trivial since everything commutes. The final case is a triple of the form \{P_{i\ell},P_{jk},P_{ij}\} as in

Here we have use the definition of the D’s and the fact that P’s with disjoint indices commute. Lemma 2.9 gives [P_{\ell i }, D_{i kj}]= [P_{kj}, D_{j \ell i}] and so the remaining terms cancel.

For triples containing two P’s and a D, a first choice is for all three to belong to an R(3) sub-algebra. This case was addressed in the previous section, Lemma 2.2.

For the other two cases, we assume that the P’s are not identical and have all 4 indices contained in them. Without loss of generality, we take D_{ijk} and then either P_{i\ell} and P_{k \ell} or P_{i \ell} and P_{jk}. The first case was discussed above where the identity (2.11) was derived.

For the second case, we compute

The first term is computed using (2.11)

Using (2.11), the second term becomes

Thus, we have

These resulting two lines are equivalent to (2.11) and the equation in reverse order (2.9) and hence add to 0.

Thus, we have shown that the Jacobi relations for two P’s and a D are satisfied without additional constraints. The proof of higher-order relations follow similarly and are included in Appendix ?.

Relations (2.12)–(2.12) and (2.12) were shown for arbitrary indices. It remains to check the final relations (2.12) and (2.12) which have no analog when n=4.

For (2.12), we begin with the identity

to obtain

Similarly, (2.12) is obtained from