← Explorations

Counting singlets.

How a column of numbers from group theory led me to a Laguerre polynomial

In this exploration
  1. A counting question
  2. What is being counted
  3. The data
  4. Following the numbers
  5. From conjecture to a proposed theorem
  6. New, as far as I can tell
  7. Reproduce it
  8. Further reading
  9. Footnotes

A counting question

How many independent ways are there to combine quarks and antiquarks into something color-neutral? With the three colors of SU(3), three quarks combine in exactly one way, and four quarks with one antiquark in three ways. This exploration follows how a pattern in a table of numbers led me to a formula that answers the question for any number of colors, a formula I have not found in the literature.

Let me state the question in general. SU(nn) is the symmetry group of nn colors.1 A quark here is something that can carry any one of the nn colors, and an antiquark is an object carrying an anticolor. Combine n+mn+m quarks with mm antiquarks. How many linearly independent color-neutral states, or singlets, can they form? I label the answer as N(n,m)N(n,m). These words may not make sense to you now; they will be defined in the next section.

I needed these numbers for my research. In a qubit-regularized gauge theory, each link of the lattice holds a small, fixed number of states, and the gauge-invariant states at a site are built from singlets of the representations on the links that meet there. In the previous exploration, these were singlets of a few spin-½ flux ends in SU(2), and I could count them by hand. For SU(nn), a computer package such as GroupMath returns the numbers one at a time.2 For practical purposes, that is all I need. However, my curiosity demanded a neater formula for N(n,m)N(n,m) in terms of nn and mm.

The figure below lets you choose nn and mm and read off N(n,m)N(n,m), together with a closed formula that I will arrive at by the end.

Singlets of n + m quarks and m antiquarks

N(3, 2) = 11
GroupSU(3)

k = n + m = 5 quarks and m = 2 antiquarks: the singlets of 3⊗5 ⊗ 3⊗2

Closed form: m! Lm(n)(1) = 2! L2(3)(1) = 11

Inside the theorem’s range (m ≤ n + 1): the count equals the closed form

Selection rule: with 2 antiquarks, every other k with m < k < n + m (here k = 3 and 4) gives 0 singlets.

SU(3), 5 quarks and 2 antiquarks: 11 independent singlets. The closed form gives 11 too; m ≤ n + 1 is inside the theorem’s range.

Each count is the number of independent singlets in the product, computed by singlets.py; the closed form is the alternating sum of the theorem. The page reads both from that script’s output and computes neither. The dots stand for the particles only; they do not show how the singlets are built.

What is being counted

Here is the one paragraph of linear algebra I need. The color state of one quark is a vector in Cn\mathbb C^n, and each element UU of SU(nn) acts on it as the n×nn\times n matrix UU. A rule that turns group elements into matrices in this way is a representation; this one is the fundamental representation, written n\mathbf{n}. An antiquark carries the dual representation nˉ\bar{\mathbf{n}}, in which UU acts by its complex conjugate Uˉ\bar U.3 The states of kk quarks and mm antiquarks form the tensor product n⊗k⊗nˉ⊗m\mathbf{n}^{\otimes k}\otimes\bar{\mathbf{n}}^{\otimes m}, a space of dimension nk+mn^{k+m}, on which UU acts as UU on every quark and as Uˉ\bar U on every antiquark. A singlet is a vector that every UU in SU(nn) leaves fixed. The singlets form a subspace, and its dimension is the number of independent color-neutral combinations.4

Not every number of quarks allows a singlet. The matrix ωI\omega I, with ωn=1\omega^n=1, belongs to SU(nn), and it multiplies every vector of n⊗k⊗nˉ⊗m\mathbf{n}^{\otimes k}\otimes\bar{\mathbf{n}}^{\otimes m} by the same number, ωk−m\omega^{k-m}.5 A singlet must be left fixed, so singlets exist only when k−mk-m is a multiple of nn. When quarks outnumber antiquarks by at least one and at most nn, that is, for m<k≤n+mm\lt k\le n+m, only k=n+mk=n+m survives. That is the case I count:

N(n,m):=N(n⊗(n+m)⊗nˉ⊗m).N(n,m):=N\bigl(\mathbf{n}^{\otimes(n+m)}\otimes\bar{\mathbf{n}}^{\otimes m}\bigr).

For SU(3), N(3,0)N(3,0) counts the singlets of three quarks, and N(3,1)N(3,1) those of four quarks and an antiquark.

The data

I first computed these counts with GroupMath. The table shows N(n,m)N(n,m) for small nn and mm.6 I have left out the cells with m>n+1m>n+1, where the pattern below breaks down.

n \ m012345
1111
212514
3131147225
414191046414364
51529191140511389
61641314266124694
71755479457747579

The n=3n=3 row begins with three numbers I had already worked out for SU(3): one singlet for three quarks, three for four quarks and an antiquark, and eleven for five quarks and two antiquarks. Read along a row, and the numbers grow quickly with no rule in sight. Read down a column instead. Fix mm, let nn grow from n=m−1n=m-1, and the entries follow a polynomial in nn. The first six are

P0(n)=1,P1(n)=n,P2(n)=n2+n−1,P3(n)=n3+3n2−n−4,P4(n)=n4+6n3+5n2−16n−15,P5(n)=n5+10n4+25n3−20n2−111n−56.\begin{aligned} P_0(n)&=1,\\ P_1(n)&=n,\\ P_2(n)&=n^2+n-1,\\ P_3(n)&=n^3+3n^2-n-4,\\ P_4(n)&=n^4+6n^3+5n^2-16n-15,\\ P_5(n)&=n^5+10n^4+25n^3-20n^2-111n-56. \end{aligned}

I fitted these polynomials from the counts up to P10P_{10}. What single formula produces all of them?

Following the numbers

Since I am not Ramanujan and I haven’t memorized enough special functions, the pattern in those polynomials is not immediately obvious to me. What follows more or less follows my original steps of discovery that led me to a conjecture: take differences, notice a pattern, build a generating function.

Differences

Differences are the discrete version of derivatives. A polynomial of degree mm becomes a constant after mm differences, so repeated differences reveal the degree of a column without any guess at its formula. Take m=3m=3 and n=2,…,8n=2,\dots,8. Subtract each entry from the next, do the same to the new row, and repeat. In the table, each entry is the number above and to its right minus the number directly above it; for instance, 33=47−1433=47-14.

Rown = 2345678
Δ0\Delta^01447104191314479692
Δ1\Delta^1335787123165213
Δ2\Delta^22430364248
Δ3\Delta^36666

The third differences are constant, and the constant is 3!=63!=6. So the m=3m=3 column comes from a cubic, and the constant 3!3! means its leading coefficient is 1.7 Newton’s forward-difference formula then reads P3P_3 off the first entry of each row,

P3(n)=14+33(n−21)+24(n−22)+6(n−23)=n3+3n2−n−4.P_3(n)=14+33\binom{n-2}{1}+24\binom{n-2}{2}+6\binom{n-2}{3}=n^3+3n^2-n-4.

The same happened in every column up to m=10m=10: the mm-th differences are the constant m!m!, so PmP_m has degree mm and leading coefficient 1.

A pattern in the first differences

The first differences hold more than the degree. Look at them again: 33, 57, 87, 123, 165, 213. Every one is a multiple of 3, and dividing by 3 gives 11, 19, 29, 41, 55, 71. That is the m=2m=2 column, starting at n=3n=3. In other words, N(n+1,3)−N(n,3)=3 N(n+1,2)N(n+1,3)-N(n,3)=3\,N(n+1,2). The same holds in every column, with the factor mm in place of 3. For instance, N(5,4)−N(4,4)=1405−641=764=4×191=4 N(5,3)N(5,4)-N(4,4)=1405-641=764=4\times191=4\,N(5,3). As an identity between polynomials, for every m≥1m\ge1,

Pm(n+1)−Pm(n)=m Pm−1(n+1).P_m(n+1)-P_m(n)=m\,P_{m-1}(n+1).

I will call this pattern D, for differences. Who would have thought one column of the table knows so much about its neighbor!

The figure below lets you repeat both steps. Choose mm and use the Take differences button until the rows collapse to the constant m!m!. Then turn on the Compare with m × previous column toggle to see pattern D.

Differences down a column

m = 3Constant: m! = 6
Differences taken3 of 3
Differences of N(n, 3) for n = 2 to 8
n2345678
Δ⁰N(n, 3)1447104191314479692
Δ¹335787123165213
3 × N(n + 1, 2)m × previous column333 × 11573 × 19873 × 291233 × 411653 × 552133 × 71
Δ²2430364248
Δ³m! = 66666

Δ³ subtracts neighboring entries of Δ²: 4 entries. Δ³ is constant: every entry is 6 = 3!. So on this window N(n, 3) is a polynomial in n of degree 3 with leading coefficient 1.

Beneath Δ¹: 3 × N(n + 1, 2), the m = 2 column taken one step along in n. All 6 entries agree with Δ¹: they match, as pattern D predicts.

Each row lists the differences of neighboring entries in the row above, so its entries sit between the two they come from. The first row is read from the output of singlets.py; the page computes only these subtractions and the products m × N in the comparison row. The window starts at n = m − 1 (n = 1 for m = 1), where the column follows its polynomial.

A generating function

A factor of mm that appears each time mm steps down by one is the signature of an exponential generating function, a power series whose mm-th term is divided by m!m!. So for each nn I packed the whole sequence P0(n),P1(n),…P_0(n),P_1(n),\dots into one function,

Gn(t):=∑m≥0Pm(n) tmm!.G_n(t):=\sum_{m\ge0}P_m(n)\,\frac{t^m}{m!}.

Multiply pattern D by tm/m!t^m/m! and sum over m≥1m\ge1. The factor mm cancels against the factorial, and the whole pattern becomes one line, Gn+1(t)−Gn(t)=t Gn+1(t)G_{n+1}(t)-G_n(t)=t\,G_{n+1}(t). Hence

Gn(t)=G0(t)(1−t)n.G_n(t)=\frac{G_0(t)}{(1-t)^n}.

But what is G0(t)G_0(t)? Here comes the exciting part. If pattern D holds, every polynomial in the family is fixed by a single sequence of integers, the constant terms:8

Pm(0)=1, 0, −1, −4, −15, −56, −185, −204, 6209, 112400, 1520271, …P_m(0)=1,\ 0,\ -1,\ -4,\ -15,\ -56,\ -185,\ -204,\ 6209,\ 112400,\ 1520271,\ \dots

A quick search on the On-Line Encyclopedia of Integer Sequences (OEIS) identified this sequence as A009940: m! Lm(1)m!\,L_m(1), where LmL_m is the mm-th Laguerre polynomial. Its exponential generating function is G0(t)=e−t/(1−t)/(1−t)G_0(t)=e^{-t/(1-t)}/(1-t), and so

Gn(t)=e−t/(1−t)(1−t)n+1.G_n(t)=\frac{e^{-t/(1-t)}}{(1-t)^{n+1}}.

Laguerre

I recognized the shape of this function (or, rather, now it is easier to look through the list of well-known generating functions). The generalized Laguerre polynomials Lm(α)(x)L_m^{(\alpha)}(x), a classical family of orthogonal polynomials, have the generating function9

∑m≥0Lm(α)(x) tm=e−xt/(1−t)(1−t)α+1.\sum_{m\ge0}L_m^{(\alpha)}(x)\,t^m=\frac{e^{-xt/(1-t)}}{(1-t)^{\alpha+1}}.

At α=n\alpha=n and x=1x=1 this matches Gn(t)G_n(t) term by term, so Pm(n)/m!=Lm(n)(1)P_m(n)/m!=L_m^{(n)}(1). That was my conjecture:

N(n,m)=Pm(n)=m! Lm(n)(1).\boxed{N(n,m)=P_m(n)=m!\,L_m^{(n)}(1).}

That was the arduous but exciting short hike I took to reach the conjecture. Not to rob others of the same fun, but I thought I could contribute to the mathematical community by saving people some time: just as I first searched the coefficients on OEIS, I wished I didn’t have to go through A009940 to get to my generating function.
Thus, in March 2024, I submitted the coefficients of nn in my polynomials, 0,1,1,−1,−16,−111,−691,…0,1,1,-1,-16,-111,-691,\dots, to the OEIS. The sequence is now A371114, m!m! times the coefficient of xx in Lm(x)(1)L_m^{(x)}(1). The formula listed there, in terms of A009940, is exactly what the generating function predicts.10

Still, the data stop at m=10m=10 and small nn. Neither pattern D nor the conjecture was proven, and I could not see why the polynomials stop matching the counts once mm exceeds n+1n+1.

And then AI enters the picture.

From conjecture to a proposed theorem

I did not find the proof myself: it was written by Claude, an AI model, in October 2026. To be honest, I have not checked it step by step. In my foreseeable future, my interest in physics will occupy so much of my attention that I won’t have time to check it. Nevertheless, I publish it as a tentative proof because I believe putting the idea to the public is better than letting it stay on my computer. If the proof holds, it turns the conjecture into the following theorem.

Theorem (proof by Claude, not yet independently checked). For every n≥1n\ge1 and 0≤m≤n+10\le m\le n+1,

N(n,m)=∑i=0m(−1)i(mi)(n+m)!(n+i)!=m! Lm(n)(1),N(n,m)=\sum_{i=0}^{m}(-1)^i\binom{m}{i}\frac{(n+m)!}{(n+i)!}=m!\,L^{(n)}_m(1),

and the other values of kk in the window m<k≤n+mm\lt k\le n+m give no singlets at all.

The script at the end of this article confirms the formula numerically for every n≤15n\le15 and m≤n+1m\le n+1, independently of the proof; the proof, if it holds, extends it to all nn.

The alternating sum shows directly that Pm(n)P_m(n) is a polynomial in nn of degree mm, with integer coefficients and leading coefficient 1, as the differences suggested. The generating function Gn(t)G_n(t) follows from the theorem for every mm, and pattern D follows from GnG_n. So if the proof holds, the pattern and the generating function are theorems too.

The range m≤n+1m\le n+1 cannot be extended while fixing the rest of the theorem. At m=n+2m=n+2, the count exceeds the formula by exactly the Catalan number Cat(n+1)=(2n+2n+1)/(n+2)\mathrm{Cat}(n+1)=\binom{2n+2}{n+1}/(n+2). For SU(2), for instance, N(2,4)=37+5=42N(2,4)=37+5=42. This is why the table stops at m=n+1m=n+1. The formula does not fail at random: just outside its range it misses by a well-known sequence, 2,5,14,42,132,4292, 5, 14, 42, 132, 429 for n=1n=1 to 66, and the proof below traces the miss to a single Young diagram. The first figure shows this excess when you set m=n+2m=n+2.

The proposed proof, in outline

Write K=n+mK=n+m and V=CnV=\mathbb C^n. A Young diagram λ⊢K\lambda\vdash K is a partition of KK, drawn as left-justified rows of boxes, and fλf^\lambda is the number of standard Young tableaux of shape λ\lambda: fillings of the boxes with 1,…,K1,\dots,K that increase along each row and down each column. The proof turns the count into a sum over such tableaux and then collapses the sum.

(i) Schur–Weyl duality. Singlets in V⊗K⊗V∗⊗mV^{\otimes K}\otimes V^{*\otimes m} are the SU(nn)-maps from V⊗mV^{\otimes m} to V⊗KV^{\otimes K}. Since ΛnV\Lambda^nV, the fully antisymmetric part of V⊗nV^{\otimes n}, is trivial for SU(nn), one may replace V⊗mV^{\otimes m} by ΛnV⊗V⊗m\Lambda^nV\otimes V^{\otimes m}, and the maps are then the same as U(n)-maps. Schur–Weyl duality splits V⊗KV^{\otimes K} into irreducible representations of U(n), each paired with an irreducible representation of the symmetric group SKS_K and labeled by a Young diagram with at most nn rows. Schur’s lemma then reduces NN to a question about the symmetric group.

(ii) Kostka numbers. After transposing the diagrams, that question has the answer

N=∑μ⊢K, μ1≤nfμ Kμ,(n,1m),N=\sum_{\mu\vdash K,\ \mu_1\le n}f^\mu\,K_{\mu,(n,1^m)},

where the Kostka number KμαK_{\mu\alpha} is the number of semistandard tableaux of shape μ\mu and content α\alpha. Two standard facts do the rest. For content (a,1K−a)(a,1^{K-a}), the aa ones must fill the start of row 1, so Kμ,(a,1K−a)=fμ/(a)K_{\mu,(a,1^{K-a})}=f^{\mu/(a)}, the number of standard fillings of the skew shape left after removing the first aa boxes of row 1. And ∑μfμKμ,(a,1K−a)=K!/a!\sum_\mu f^\mu K_{\mu,(a,1^{K-a})}=K!/a!, the dimension of a permutation module of SKS_K.

(iii) The sieve lemma. For μ⊢n+m\mu\vdash n+m with m≤n+1m\le n+1,

∑i=0m(−1)i(mi)fμ/(n+i)=[μ1≤n] fμ/(n),\sum_{i=0}^m(-1)^i\binom{m}{i}f^{\mu/(n+i)}=[\mu_1\le n]\,f^{\mu/(n)},

where [P][P] is 1 if PP holds and 0 otherwise. If μ1≤n\mu_1\le n, only the i=0i=0 term survives. If μ1=n+r\mu_1=n+r with r≥1r\ge1, then μ1+μ2≤n+m≤2n+1\mu_1+\mu_2\le n+m\le 2n+1 forces μ2≤n\mu_2\le n. This is the only place where m≤n+1m\le n+1 is used. The skew shape then splits into a single row of r−ir-i boxes and the rest of the diagram, and the left side becomes a multiple of (mr)∑i=0r(−1)i(ri)=0\binom{m}{r}\sum_{i=0}^r(-1)^i\binom{r}{i}=0.

(iv) The alternating sum. Substituting the lemma into the Kostka sum, swapping the two sums and using both Kostka facts gives

N=∑i=0m(−1)i(mi)(n+m)!(n+i)!.N=\sum_{i=0}^m(-1)^i\binom{m}{i}\frac{(n+m)!}{(n+i)!}.

(v) Laguerre. The explicit form of Lm(n)(1)L_m^{(n)}(1) turns this sum into m! Lm(n)(1)m!\,L_m^{(n)}(1). Equivalently, the alternating sum has exactly the generating function Gn(t)G_n(t) found from the data.

At m=n+2m=n+2, the condition μ2≤n\mu_2\le n fails for exactly one diagram, μ=(n+1,n+1)\mu=(n+1,n+1), two rows of n+1n+1 boxes. For it the left side of the lemma is −1-1, while the right side is 0, so the count picks up an extra f(n+1,n+1)=Cat(n+1)f^{(n+1,n+1)}=\mathrm{Cat}(n+1).

There is also a heuristic reading of the theorem. The i=0i=0 term, (n+m)!/n!=(n+mn) m!(n+m)!/n!=\binom{n+m}{n}\,m!, counts an obvious spanning set of singlets: put one ε\varepsilon-tensor, the fully antisymmetric tensor with nn indices, on nn of the quarks, and pair each of the remaining mm quarks with an antiquark by a Kronecker delta. These singlets are linearly dependent, and the alternating terms remove the overcounting.

New, as far as I can tell

The ingredients are all on record: the Laguerre polynomials and their generating function are classical, and the constant terms were already in the OEIS. What I have not found in the literature is the statement that connects them to singlet counts, N(n,m)=m! Lm(n)(1)N(n,m)=m!\,L_m^{(n)}(1) for m≤n+1m\le n+1. The formula is new to me, and I would be glad to learn of an earlier source.

The same singlet counts N(n,m)N(n,m) can also be written as integrals over the group,11

N(n,m)=∫SU(n)dU (tr⁡U)n+m(tr⁡U†)m,N(n,m)=\int_{SU(n)}dU\,(\operatorname{tr}U)^{n+m}(\operatorname{tr}U^\dagger)^m,

where dUdU is the Haar measure, the uniform measure on the group, normalized so that the group has total volume 1. The Weyl integration formula turns this into an integral over the eigenvalues of UU. The same counts give the number of independent tensor structures in the group integrals ∫dU U⊗(N+m)⊗U†⊗m\int dU\,U^{\otimes(N+m)}\otimes U^{\dagger\otimes m} of the kind Zuber revisited for SU(NN). Such integrals appear in strong-coupling expansions of baryonic Wilson loops.

Now I will say something stupid that may frustrate a mathematician: the proof indirectly proves/derives pattern D, but does not explain why. Is there a bijection, a matching of the singlets or of the tableaux that count them, that shows directly why N(n+1,m)−N(n,m)N(n+1,m)-N(n,m) equals mm times N(n+1,m−1)N(n+1,m-1)? Such a bijection would explain the pattern that started this search in terms of the singlets themselves. How can I “see” this fact? Perhaps the tableau sum is where I would start.12

Reproduce it

The script I used is singlets.py. It needs only Python 3, with no extra packages, and python3 singlets.py finishes in under a second. It checks eight things:

  1. a fast formula for N(n,m)N(n,m) as a sum over Young tableaux, against a direct evaluation of the group integral, together with the selection rule;
  2. the data table;
  3. the difference table for m=3m=3;
  4. the polynomials P0P_0 to P10P_{10}, fitted from the counts and checked against their expected coefficients;
  5. pattern D;
  6. the constant terms against A009940, and the generating function;
  7. the theorem for every n≤15n\le 15 and m≤n+1m\le n+1;
  8. the Catalan excess at m=n+2m=n+2.

Further reading

  • J.-B. Zuber, “Revisiting SU(N) integrals,” Journal of Physics A: Mathematical and Theoretical 50, 015203 (2017), arXiv:1611.00236. It studies SU(NN) group integrals whose independent tensor structures these counts enumerate.
  • R. M. Fonseca, “GroupMath: A Mathematica package for group theory calculations”, Computer Physics Communications, 108085 (2021), arXiv:2011.01764. The package I used for the original counts.
  • W. Fulton and J. Harris, Representation Theory: A First Course (Springer, 1991), Lectures 4 and 6. A standard reference for Schur–Weyl duality, Young tableaux and Kostka numbers.
  • D. Bump, Lie Groups, 2nd ed., Graduate Texts in Mathematics 225 (Springer, 2013), Chapters 2 and 17. A standard reference for Schur orthogonality of characters, which writes the counts as group integrals, and for the Weyl integration formula.
  • DLMF, Eq. 18.12.13. The generating function of the Laguerre polynomials.
  • OEIS, A009940: a(n)=n! Ln(1)a(n)=n!\,L_n(1), the constant terms of the polynomials.
  • R. X. Siew, OEIS, A371114: n!n! times the coefficient of xx in Ln(x)(1)L_n^{(x)}(1), the linear coefficients of the polynomials.

Footnotes

  1. SU(nn) is the group of n×nn\times n unitary matrices with determinant 1. The quarks and antiquarks are names for its two simplest representations, defined in the next section. Only for n=3n=3 are they the quarks of particle physics. ↩

  2. GroupMath, by R. M. Fonseca, is a Mathematica package for group theory calculations; see Further reading. ↩

  3. For a unitary matrix, Uˉ=(U−1)T\bar U=(U^{-1})^{\mathsf T}, which is how UU acts on linear functions of a vector. That is why the antiquark representation is called the dual. Writing V=CnV=\mathbb C^n, the antiquarks live in the dual space V∗V^*. ↩

  4. Equivalently, the singlets in V⊗k⊗V∗⊗mV^{\otimes k}\otimes V^{*\otimes m} are the linear maps from V⊗mV^{\otimes m} to V⊗kV^{\otimes k} that commute with every UU. In index language, NN counts the linearly independent invariant tensors with kk fundamental and mm antifundamental indices. ↩

  5. These matrices form the center of SU(nn), the elements that commute with every other element. Each quark contributes a factor ω\omega and each antiquark a factor ωˉ=ω−1\bar\omega=\omega^{-1}. ↩

  6. The values shown are computed by the script linked at the end of this article. It does not use GroupMath, and its SU(3) counts agree with the original GroupMath values. ↩

  7. For a polynomial of degree mm with leading coefficient aa, the mm-th difference is the constant m! am!\,a. Newton’s formula P(n)=∑jΔjP(n0)(n−n0j)P(n)=\sum_j \Delta^jP(n_0)\binom{n-n_0}{j} rebuilds the polynomial from the first entry ΔjP(n0)\Delta^jP(n_0) of each row. ↩

  8. Here n=0n=0 does not correspond to a special unitary group; Pm(0)P_m(0) is simply the polynomial evaluated at zero. ↩

  9. Explicitly, Lm(α)(x)=∑i=0m(−1)i(m+αm−i)xii!L_m^{(\alpha)}(x)=\sum_{i=0}^m(-1)^i\binom{m+\alpha}{m-i}\frac{x^i}{i!}. The generating function is given in the NIST Digital Library of Mathematical Functions (DLMF), Eq. 18.12.13, and is valid for ∣t∣<1|t|\lt 1. The ordinary Laguerre polynomials LmL_m are the case α=0\alpha=0. ↩

  10. The formula is a(m)=m!∑k=0m−1A009940(k)/(k! (m−k))a(m)=m!\sum_{k=0}^{m-1}\mathrm{A009940}(k)/\bigl(k!\,(m-k)\bigr), the coefficient form of ∂nGn(t)∣n=0=−ln⁡(1−t) G0(t)\partial_nG_n(t)\big|_{n=0}=-\ln(1-t)\,G_0(t). ↩

  11. The integrand (tr⁡U)k(tr⁡U†)m(\operatorname{tr}U)^k(\operatorname{tr}U^\dagger)^m is the trace of the matrix by which UU acts on n⊗k⊗nˉ⊗m\mathbf{n}^{\otimes k}\otimes\bar{\mathbf{n}}^{\otimes m}. Averaging that matrix over the group projects every vector onto its singlet part, and the trace of this projection is the number of singlets. This is Schur orthogonality for the characters of a compact group; see Bump, Lie Groups, Chapter 2. The eigenvalue form is the Weyl integration formula, Chapter 17 of the same book. ↩

  12. A natural place to look is the tableau sum N(n,m)=∑ν⊢m, ℓ(ν)≤nfν fν+(1n)N(n,m)=\sum_{\nu\vdash m,\ \ell(\nu)\le n}f^{\nu}\,f^{\nu+(1^n)}, valid for every m≥0m\ge0, where ℓ(ν)\ell(\nu) is the number of rows of ν\nu and ν+(1n)\nu+(1^n) adds a full column of height nn to ν\nu. The question is then whether some bijection between pairs of standard Young tableaux accounts for pattern D. ↩