Abstract
We demonstrate how formulas that express Hecke-type double-sums in terms of theta functions and Appell–Lerch functions—the building blocks of Ramanujan’s mock theta functions—can be used to give general string function formulas for the affine Lie algebra \(A_{1}^{(1)}\) for levels \(N=1,2,3,4\).
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Notation
Let q be a complex number where \(q:=e^{2\pi i \tau }\) and \(\tau \in \mathfrak {H}:=\{z\in \mathbb {C} | \Im {z}>0\}\). We recall basic notation such as
where in the last line the equivalence of product and sum follows from Jacobi’s triple product identity. We draw the reader’s attention to the fact that \(j(q^n;q)=0\) for \(n\in \mathbb {Z}.\)
Let a and m be integers with m positive. We give special notation to frequently encountered theta functions
and we also recall Dedekind’s eta-function:
We define an Appell–Lerch function as follows. Let \(x,z\in \mathbb {C}\backslash \{0\}\) with neither z nor xz an integral power of q. Then
Lastly, we recall a useful form of a Hecke-type double-sum [6]. For \(x,y\in \mathbb {C}\backslash \{0\}\),
2 Introduction
In [10], Kac and Peterson give several examples of string functions for affine Lie algebras of type \(A_{1}^{(1)}\) that have beautiful evaluations in terms of theta functions. See also [11, 12]. Their string functions are closely related to the real quadratic fields \(\mathbb {Q}(\sqrt{m(m+2)})\). Indeed, if we fix a positive integer m, their string functions are of the form [10, p. 260]:
where N and n are integers with \(n\equiv N \pmod 2\). Here we have replaced Kac and Peterson’s notation (n, N, m) with \((m,\ell ,N)\) of [13]. In this form, m and \(\ell \) parametrize the maximal (resp. highest) weight in terms of the fundamental weights of the affine Kac–Moody algebra \(\mathfrak {g}=A_1^{(1)}\). See [10] and [13] for more details on string functions.
From [13] we recall that \(m,\ell , N\) are integers with \(N\ge 1\), \(\ell \in \{0,1,2,\dots ,N \}\), and \(m\equiv \ell \pmod 2\). From [10, p. 260], [13] we find that
where
and
In [9] we derived the useful form:
Identity (2.5) is very useful in computing the modularity of string functions. In [6], one finds many formulas where Hecke-type double-sums are expressed in terms of Appell-Lerch functions and theta functions. Appell–Lerch functions are the building blocks of Ramanujan’s mock theta functions. A simple example of general results found in [6] reads
Such double-sum formulas provide a straightforward method for proving identities for Ramanujan’s mock theta functions; in particular, the formulas give new proofs of the mock theta conjectures [4,5,6]. As an example, one sees from (2.6) that
where we remind the reader that \(j(x;q)=0\) if and only if x is an integral power of q.
String functions satisfy many symmetries [13]:
Looking for a computationally easier approach to using (2.5) and formulas found in [6] led to new symmetries. The author, Postnova, and Solovyev found [9]:
Theorem 2.1
[9, Theorem 1.1] We have
Corollary 2.2
[9, Corollary 1.2] We have
Corollary 2.3
[9, Corollary 1.3] For \(K\equiv \ell \pmod 2\), we have
In this paper, we will use the relation (2.5), the double-sum formulas of [6], and the new string function symmetries of [9] to give general string functions identities. In particular we will prove general string function identities for levels \(N=1,2,3,4\):
Theorem 2.4
For \(\ell \in \{0,1\}\) and \(m\equiv \ell \pmod 2\), we have
Theorem 2.5
For \(\ell \in \{0,1,2\}\), \(0\le m < 4\), \(m\equiv \ell \pmod 2\), we have
Theorem 2.6
For \(\ell \in \{0,1,2, 3\}\), \(0\le m < 6\), \(m\equiv \ell \pmod 2\), we have
where
Theorem 2.7
For \(\ell \in \{0,1,2, 3, 4\}\), \(0\le m < 8\), and \(m\equiv \ell \pmod 2\), we have
where
We then demonstrate how the general string function identities give as special cases examples found in [10, p. 220]:
Level 2:
Level 3:
Level 4:
As an interesting consequence, once one has Theorems 2.4, 2.5, 2.6, 2.7 in mind, identities such as (2.20a), (2.21b), (2.21c), and (2.22a) become special cases of the classic theta function identity:
In particular identities (2.20a), (2.21b), (2.21c), and (2.22a) follow from the specializations \(m=2,3,3,12\) respectively.
In Section 3, we recall background information on theta functions, Appell–Lerch functions, and Hecke-type double-sums. In Section 4 we present a proof of Theorem 2.4. The proof is a corrected version of a sketch found in [6, Example 1.3]. In Section 5 we present a new proof of Theorem 2.5. In Section 6 we present a new proof of Theorem 2.6. In Section 7 we present a new proof of Theorem 2.7. In Section 8, we use (2.15) to prove the level \(N=2\) identities. In Section 9, we use (2.16) to prove the level \(N=3\) identities. In Section 10, we use (2.18) to prove the level \(N=4\) identities.
Although there are general formulas for level N string functions, see for example [7] and [13, (6.5), (6.6)], we emphasize that our methods here are new. In particular, Kac and Peterson appeal to modularity to prove the string function identities [10, p. 220]. They employ the transformation law for string functions under the full modular group, calculate the first few terms in the Fourier expansions of the string functions, and exploit the fact that a modular form vanishing at cusps to sufficiently high order is zero. In the present paper, we use the relation (2.5), the double-sum formulas of [6].
We point out that the formula for \(N=1\) is well-known, and the formula for \(N=2\) is related to the Ising model in statistical mechanics. Our formulations for \(N=3\) and \(N=4\) appear to be new; however, some of the pieces can be found in [10, pp. 219-220].
3 Preliminaries
3.1 Theta functions
We collect some frequently encountered product rearrangements:
Following from the definitions are the following general identities:
if \(n\ge 1\), \(\zeta _n\) is a primitive n-th root of unity.
A convenient form of the Weierstrass three-term relation for theta functions is,
Proposition 3.1
For generic \(a,b,c,d\in \mathbb {C}^*\)
We collect several useful results about theta functions in terms of a proposition [1, 4, 5]:
Proposition 3.2
For generic \(x,y,z\in \mathbb {C}^*\)
We finish this subsection with a series of lemmas.
Lemma 3.3
We have
Proof
We specialize (3.2d) to obtain
Hence
\(\square \)
Lemma 3.4
We have
Proof (Proof of Lemma 3.4)
From (2.23) with \(m=12\), we have
Substituting \(x\rightarrow q^{1/2}\) and using identities (3.1a) and (3.1b) yields
The result follow from identity (3.1b) and the substitution \(q\rightarrow q^{1/6}\). \(\square \)
3.2 Appell–Lerch functions
The Appell-Lerch function satisfies several functional equations and identities [6, 15]:
Proposition 3.5
For generic \(x,z\in \mathbb {C}^*\)
Corollary 3.6
We have
3.3 Hecke-type double-sums
We recall a few basic properties of Hecke-type double-sums. We have a proposition and a corollary:
Proposition 3.7
[6, Proposition 6.3] For \(x,y\in \mathbb {C}\backslash \{0\}\) and \(R,S\in \mathbb {Z}\)
We also have the property [6, (6.2)]:
In order to state the double-sum formulas that we will be using, we introduce the useful
In [6, Theorem 1.3], we specialize \(n=1\), to have
Theorem 3.8
Let p be a positive integer. For generic \(x,y\in \mathbb {C}^*\)
where
The specialization for \(p=1\) will be of importance. It is just (2.6):
Corollary 3.9
We have
For another useful result, we specialize [6, Theorem 1.4] to \(a=b=n\), \(c=1\).
Theorem 3.10
Let n be a positive integer. Then
where
and
Theorem 3.10 has the following specializations.
Corollary 3.11
We have
where
Corollary 3.12
We have
where
In Theorem 3.8, we set \(z_1=z_0=-1\) in the Appell–Lerch expression (3.10). For examples where \(p=2,3\), we can set \(z_1=z_0^{-1}=y/x\) to reduce the number of theta quotients. For example, we can specialize [6, Theorem 1.9] to \(n=1\) to have
Theorem 3.13
For generic \(x,y\in \mathbb {C} \backslash \{ 0\}\)
where
We can also specialize [6, Theorem 1.10] to \(n=1\) to have
Theorem 3.14
For generic \(x,y\in \mathbb {C} \backslash \{ 0\}\)
where
We can also specialize [6, Theorem 1.11] to \(n=1\) to have
Theorem 3.15
Let n be a positive odd integer. For generic \(x,y\in \mathbb {C} \backslash \{ 0\}\)
where
and
with
Proposition 3.16
[6, Proposition 8.1] Let \(\ell \in \mathbb {Z}\), \(p\in \{ 1,2,3,4\}\), \(n\in \mathbb {N}\) with \((n,p)=1\). For generic \(x,y\in \mathbb {C} \backslash \{ 0\}\)
3.4 The general integral-level string function
We recall the notation
and the fact that [13]:
The following is a straightforward consequence of the symmetry relations (2.8) - (2.10).
Lemma 3.17
If
then
4 Computing the general level \(N=1\) string function
Let us set \(N=1\). Here \(\ell \in \{0,1\}\), \(0\le m<2\), \(m\equiv \ell \pmod 2\). The proof of Theorem 2.4 follows from a lemma, whose proof we do now.
Lemma 4.1
We have
Proof of Lemma 4.1
This is just the calculation found in (2.7). \(\square \)
Proof of Theorem 2.4
For \(N=1\) we only need to be concerned with \(\ell \in \{0,1\}\), \(0\le m <2\), \(m\equiv \ell \pmod 2\). Which means we only need to compute the string functions for the \((\ell ,m)\) tuples (0, 0) and (1, 1). For the case \((\ell ,m)=(0,0)\), we have
where the last equality follows from Lemma 4.1. The case \((\ell ,m)=(1,1)\) follows from Lemma 3.17. \(\square \)
5 Computing the general level \(N=2\) string function
Here \(N=2\), \(\ell \in \{0,1,2\}\), \(0\le m < 4\), and \(m\equiv \ell \pmod 2\). We begin with a proposition.
Proposition 5.1
We have
Proof of Proposition 5.1
We use Theorem 3.13 and Proposition 3.16 with the specialization \(n=1\):
We prove (5.1). We use (5.4) with \(k=1\) to have
The m(x, z; q) terms are defined, and their theta coefficients are both zero. Hence
where we have used (3.1a), (3.1b), and elementary product rearrangements.
We prove (5.2). We use (5.4) with \(k=0\) to have
where we have used (3.1a), (3.1b), and elementary product rearrangements.
We prove (5.3). We use (5.4) with \(k=1\) to have
where we have used (3.1a), (3.1b), and elementary product rearrangements. \(\square \)
Proof of Theorem 2.5
For \(N=2\) we only need to be concerned with \(\ell \in \{0,1, 2\}\), \(0\le m <4\), \(m\equiv \ell \pmod 2\). Which means we only need to compute the string functions for the \((\ell ,m)\)-tuples (0, 0), (0, 2), (1, 1), (1, 3), (2, 0), (2, 2). Because of Lemma 3.17, we only need to be concerned with the \((\ell ,m)\)-tuples (0, 0), (1, 1), (2, 0).
For \((\ell ,m)=(0,0)\), we have
where the last equality follows from identity (5.1). For \((\ell ,m)=(1,1)\), we have
where the last equality follows from identity (5.2). For \((\ell ,m)=(2,0)\), we have
where the last equality follows from identity (5.3). \(\square \)
6 Computing the general level \(N=3\) string function
Here \(N=3\), \(\ell \in \{0,1,2, 3\}\), \(0\le m < 6\), and \(m\equiv \ell \pmod 2\). The proof of Theorem 2.6 follows from a proposition.
Proposition 6.1
We have
Proof of Proposition 6.1
We use Theorem 3.14 and Proposition 3.16:
We prove (6.1). In (6.5), set \(k=1\)
where we have used (3.1a). Simplifying, we have
where we have twice used the product rearrangement \(J_{1,5}J_{2,5}=J_1J_5\).
For (6.2), we recall (6.5) and set \(k=0\) to have
where for the last equality we used (3.1a), (3.1b), and elementary product rearrangements.
We prove (6.3). In (6.5), we set \(k=2\):
For (6.4), we take (6.5) and set \(k=1\):
\(\square \)
Proof of Theorem 2.6
For \(N=3\) we only need to be concerned with \(\ell \in \{0,1, 2, 3\}\), \(0\le m <6\), \(m\equiv \ell \pmod 2\). Because of Lemma 3.17, we only need to be concerned with the \((\ell ,m)\)-tuples (0, 0), (1, 1), (2, 0), (3, 1).
For \((\ell ,m)=(0,0)\), we have
where the last equality follows from identity (6.1). For \((\ell ,m)=(1,1)\), we have
where the last equality follows from identity (6.2). For \((\ell ,m)=(2,0)\), we have
where the last equality follows from identity (6.3). For \((\ell ,m)=(3,1)\), we have
where the last equality follows from identity (6.4). \(\square \)
7 Computing the general level \(N=4\) string function
Proof of Theorem 2.7
For \(N=4\) we only need to be concerned with \(\ell \in \{0,1, 2, 3, 4\}\), \(0\le m <8\), \(m\equiv \ell \pmod 2\). Because of Lemma 3.17, we only need to be concerned with the \((\ell ,m)\)-tuples (0, 0), (0, 4), (0, 2), (1, 1), (1, 3), (2, 0), (2, 2). For each \((\ell ,m)\)-tuple, one uses Proposition 7.2 to compute
\(\square \)
Proposition 7.1
We have
Proposition 7.2
We have
Proof of Proposition 7.1
We recall
We prove identities (7.2a) and (7.2b). From [9, Theorem 1.1], [10, p. 219], we have
where the last equality follows from (7.1a). Similarly, we obtain an identity not in [10]:
where the last equality follows from (7.1b). Hence
The two identities (7.2a) and (7.2b) then follow from (7.3).
We prove identity (7.2c). From (2.5), [10, p. 219], we have
where the last equality follows from (7.1e). Identity (7.2d) then follows from (7.3).
We prove identities (7.2d) and (7.2e). From [9, Theorem 1.1], [10, p. 220], we have
where the last equality follows from (7.1f). Similarly, we obtain an identity not in [10]:
where the last equality follows from (7.1g). Hence
where we have used (2.23) with \(m=2\). Using (3.1b), we have
and the identities (7.2d) and (7.2e) follow from (7.3).
We prove identity (7.2f). From (2.5), we obtain an identity which is not in [10]:
where the last equality follows from (7.1c). Identity (7.2f) then follows from (7.3).
We prove identity (7.2g). Using [9, Corollary 1.3], we obtain an identity which is not in [10]:
where the last equality follows from (7.1d). Identity (7.2g) then follows from (7.3). \(\square \)
Proof of Proposition 7.1
Identity (7.1a) is true by [8, Lemma 3.11].
We prove (7.1b). We recall Corollary 3.12. The contribution from (3.15) reads
where we have used (3.1a). Hence
where for the penultimate equality we used (3.2c).
We prove (7.1c). Using (3.8) with \((R,S)=(-2,1)\), it is equivalent to show
We recall Theorem 3.15. We have
where \(x\rightarrow q^5\) and \(y\rightarrow q^{-7}\) yield
and
with
and
Hence
Using (3.2a) with \(q\rightarrow q^{12}\), \(x=y=q^2\) we have
We prove (7.1d). Here we show
We recall Corollary 3.12. The contribution from (3.15) reads
where the second equality follows from (3.5d). Hence
Continuing, we have
where the last equality follows from Lemma 3.3.
We prove (7.1e). Using (3.8) with \((R,S)=(0,1)\), it is equivalent to show
We recall Theorem 3.15. Arguing as in the proof of identity (7.1c), we find that under the substitutions \(x\rightarrow q^7\) and \(y\rightarrow q\), we have
and that
with
and
Assembling the pieces, we have
where for the last equality we used (3.1d). Using Proposition 3.1 with \(q\rightarrow q^{12}\), \(a=-q^5\), \(b=q^4\), \(c=q^2\), \(d=-i\) yields
We prove (7.1f). We recall Corollary 3.12. The contribution from (3.15) reads
where for the last equality we used (3.1a). Thus
where we have simplified using (3.1a). Regrouping terms, we have
where we used (3.2b) and (3.2c) for the second equality, regrouped terms, used elementary product rearrangements, used (3.2c) for the penultimate equality, and then finished with more product rearrangements.
We prove (7.1g). This follows from substituting \(q\rightarrow -q\) in (7.1f). \(\square \)
8 Computing level \(N=2\) string functions: Examples
8.1 The string function \(c_{20}^{20}-c_{02}^{20}\) (2.20a):
We give two proofs of identity (2.20a).
For the first proof, we use Theorem 2.6 to obain
Combining terms and using (3.1b), we have
where used (2.23) with \(m=2\) and the two product rearrangements \(J_{1,2}=J_{1}^2/J_{2}\) and \(J_{1,4}=J_{1}J_{4}/J_{2}\).
For the second proof, we use [9, Theorem 1.1] to obtain
We next recall Corollary 3.11. We have
where the last equality follows from (3.1a). Hence from Corollary 3.11:
where the last equality follows from product rearrangements.
9 Computing level \(N=3\) string functions: Examples
9.1 The string function \(c_{12}^{30}\) (2.21a):
Using Theorem 2.6 gives
9.2 The string function \(c_{30}^{30}-c_{12}^{30}\) (2.21b):
Using Theorem 2.6 gives
From (2.23) with \(m=3\) and (3.1a), we have
and under the substitution \(q\rightarrow q^{1/3}\) we have
Identity (2.21b) is now straightforward
9.3 The string function \(c_{21}^{21}-c_{03}^{21}\) (2.21c):
Using Theorem 2.6 gives
From (2.23) with \(m=3\) and (3.1a), we have
The substitution \(q\rightarrow q^{1/3}\) yields
Hence
10 Computing level \(N=4\) string functions: Examples
10.1 The string function \(c_{40}^{40}-2c_{22}^{40}+c_{04}^{40}+2c_{04}^{22}-2c_{22}^{22}\) (2.22a)
Using Theorem 2.6 gives
Hence
where the second equality follows from (2.23) with \(m=2\), and the last equality follows from Lemma 3.4.
Data Availability
Data sharing not applicable to this article as no datasets were generated or analysed during the current study.
References
Andrews, G.E., Hickerson, D.R.: Ramanujan’s “lost’’ notebook. VII: the sixth order mock theta functions. Adv. Math. 89(1), 60–105 (1991)
Bouknegt, P., Ludwig, A.W.W., Schoutens, K.: Spinon basis for high level \(SU(2)\) WZW models. Phys. Lett. B 359, 304–312 (1995)
Distler, J., Qiu, Z.: BRS cohomology and a Feigin-Fuchs representation of Kac–Moody and parafermionic theories. Nucl. Phys. B 336, 533–546 (1990)
Hickerson, D.R.: A proof of the mock theta conjectures. Invent. Math. 94(3), 639–660 (1988)
Hickerson, D.R.: On the seventh order mock theta functions. Invent. Math. 94(3), 661–677 (1988)
Hickerson, D. R., Mortenson, E. T.: Hecke-type double sums, Appell–Lerch sums, and mock theta functions, I. Proc. Lond. Math. Soc. (3) 109(2), 382–422 (2014)
Kac, V., Peterson, D.: Infinite-dimensional lie algebras, theta functions and modular forms. Adv. Math. 53, 125–264 (1984)
Kac, V., Wakimoto, M.: Modular invariant representations of infinite-dimensional Lie algebras and superalgebras. Proc. Natl. Acad. Sci. U.S.A. 85(14), 4956–4960 (1988)
Kac, V., Wakimoto, M.: Classification of modular invariant representations of affine algebras. In: Advanced Series in Mathematical Physics, vol. 7, pp. 138–177. World Scientific Publishing, Teaneck (1989)
Lepowsky, J., Primc, M.: Structure of the standard modules for the affine Lie algebra \(A_{1}^{(1)}\). In: Contemporary Mathematics, vol. 46. AMS, Providence (1985)
Mortenson, E.T.: A heuristic guide to evaluating triple-sums. Hardy-Ramanujan J. 43, 99–121 (2021)
Mortenson, E.T., Postnova, O., Solovyev, D.: On string functions and double-sum formulas. arXiv:2107.06225
Schilling, A., Warnaar, S.O.: Conjugate Bailey pairs. Contemp. Math. 197, 227–255 (2002)
Warnaar, S.O.: 50 years of Bailey’s lemma. In: Betten, A., et al. (eds.) Algebraic Combinatorics and Applications, pp. 333–347. Springer, Berlin (2001)
Zwegers, S.P.: Mock theta functions. Ph.D. Thesis, Universiteit Utrecht (2002)
Acknowledgements
We would like to thank O. Warnaar for helpful comments and suggestions.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Competing Interests
The author has no competing interests to declare that are relevant to the content of this article.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This research was supported byMinistry of Science and Higher Education of the Russian Federation, agreement No. 075-15-2019-1619.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Mortenson, E.T. On Hecke-type double-sums and general string functions for the affine Lie algebra \(A_{1}^{(1)}\). Ramanujan J 63, 553–582 (2024). https://doi.org/10.1007/s11139-023-00737-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11139-023-00737-x