---
title: Polynomial Weighted Theta Functions
url: https://www.emergentmind.com/topics/polynomial-weighted-theta-functions
type: topic
---

# Polynomial Weighted Theta Functions

Polynomial weighted theta functions are theta-type generating functions in which polynomial data modifies either the summand, the kernel, or the coefficient structure. In the classical Euclidean lattice setting, the standard form is the weighted theta series
\[
\Theta_{L,P}(q)=\sum_{v\in L}P(v)\,q^{\langle v,v\rangle/2},
\]
with \(P\) typically harmonic [1111.2392]. Closely related literatures use the same idea in non-equivalent ways: vector-valued Siegel theta functions with polynomial insertions in the Schwartz kernel [2009.06012], orthogonal-theta kernels with explicit polynomial factors depending on the Grassmannian variable [1403.3567], code- and lattice-theoretic theta series obtained by substituting theta functions into polynomial weight enumerators [1309.3812], and finite-gap tau-functions expressed as theta functions multiplied by the exponential of a quadratic polynomial rather than by a polynomial factor [1807.03377].

## 1. Terminology and conceptual range

The expression “polynomial weighted theta function” does not have a single uniform meaning across the cited literature. In some works it denotes an actual polynomial insertion into a lattice or Siegel theta kernel, as in \(\Theta_{L,P}\) or \(\Theta_L(\tau;v,p_v)\) [1111.2392] [2009.06012]. In others, the relevant structure is a theta function whose coefficients are controlled by a polynomial invariant, such as a complete or symmetric weight enumerator of a code [1309.3812]. In still other settings, the nearest analogue is not polynomial multiplication at all: for KdV tau-functions the finite-gap expression is
\[
\tau(\mathbf t)=e^{\frac12\sum_{i,j\ge0}q_{ij}t_it_j}\,
\theta\!\left(\sum t_k\mathbf V^{(k)}-\mathbf u_0\right),
\]
so the “weight” is an exponential of a quadratic polynomial [1807.03377].

This terminological spread is mirrored by modern generalizations. Scattering-diagram theta functions are weighted by tropical multiplicities defined through iterated Lie brackets rather than by a polynomial factor [1503.06183]. Weighted theta functions for non-commutative graphs use a positive semidefinite weight matrix instead of a vertex-weight vector [2101.00162]. Theta series for quantum loop algebras and Yangians decompose into root-lattice weight components that are polynomial in the spectral variable [2307.04043]. This suggests that the most stable encyclopedic meaning is umbrella-like: a polynomial weighted theta function is a theta object whose analytic, algebraic, or combinatorial content is modified by polynomial data.

## 2. Euclidean lattice theta series and harmonic polynomial weights

In the Euclidean lattice setting, the basic weighted object is
\[
\Theta_{L,P}(q)=\sum_{v\in L}P(v)\,q^{\langle v,v\rangle/2}
=\sum_{k\ge0}\left(\sum_{\langle v,v\rangle=2k}P(v)\right)q^k,
\]
where \(L\subset \mathbf R^n\) is a lattice and \(P\) is a polynomial on \(\mathbf R^n\) [1111.2392]. The unweighted theta function \(\Theta_L(q)\) is recovered by taking \(P=1\). In this form, the polynomial weight records shell statistics rather than merely shell cardinalities.

The distinguished class of weights is the space of harmonic polynomials. If \(\mathscr P_d\) denotes homogeneous degree-\(d\) polynomials and \(\mathscr P_d^0=\ker(\Delta:\mathscr P_d\to\mathscr P_{d-2})\), then
\[
\mathscr P_d=\bigoplus_{k=0}^{\lfloor d/2\rfloor}F^k\mathscr P_{d-2k}^0
=\mathscr P_d^0\oplus F\mathscr P_{d-2},
\]
with \(F=\langle x,x\rangle\) [1111.2392]. Harmonicity is structurally important because for
\[
f(x)=P(x)e^{-\pi\langle x,x\rangle t},
\]
one has the clean Fourier-transform identity
\[
\widehat f(y)=i^d\, t^{-(n/2+d)}\,P(y)e^{-\pi\langle y,y\rangle/t}
\]
when \(P\in\mathscr P_d^0\) [1111.2392]. By Poisson summation, this yields the weighted-theta functional equation
\[
\Theta_{L^*,P}(e^{-2\pi t})
= i^d\,\Vol(\mathbf R^n/L)\, t^{-(n/2+d)}\,\Theta_{L,P}(e^{-2\pi/t}).
\]

For a Type II lattice, the weighted series
\[
\theta_{L,P}(\tau)=\sum_{v\in L}P(v)e^{\pi i\langle v,v\rangle\tau}
\]
is a modular form of weight \(n/2+d\), and if \(d>0\) it is a cusp form [1111.2392]. In this form, polynomial weighting is not an auxiliary decoration but the mechanism by which shell distributions, design conditions, and modular constraints are encoded.

## 3. Siegel theta kernels and orthogonal polynomial insertions

A more general framework appears in vector-valued Siegel theta theory. For an even lattice \(L\) of signature \((b_+,b_-)\), a decomposition \(v\in \operatorname{Gr}(L_\mathbf R)\), and a polynomial \(p_v\) homogeneous of degree \((m_+,m_-)\) with respect to \(v\), the polynomially weighted Siegel theta function is
\[
\Theta_L(\tau;v,p_v)
= y^{\frac{b_-}{2}+m_-}
\sum_{\lambda\in L^*}
e\!\left(\frac{\tau\lambda_{v_+}^2}{2}+\frac{\bar\tau\lambda_{v_-}^2}{2}\right)
\bigl(e^{-\Delta_v/8\pi y}(p_v)\bigr)(\lambda)\,e_{\lambda+L},
\]
and its generalized form also depends on \(\alpha,\beta\in L_\mathbf R\) [2009.06012]. The modularity statement is explicit:
\[
\Theta_L\!\left(A\tau;A\binom{\alpha}{\beta};v,p_v\right)
=
\phi(\tau)^{\,b_+-b_-+2m_+-2m_-}\,
\rho_L(A,\phi)\,
\Theta_L\!\left(\tau;\binom{\alpha}{\beta};v,p_v\right),
\]
so for \(\alpha=\beta=0\) the weight is \(\frac{b_+-b_-}{2}+m_+-m_-\) [2009.06012].

The same paper develops a seesaw formalism for polynomially weighted theta functions. If \(L\) contains a primitive non-degenerate sublattice \(M\), \(v=u\oplus u^\perp\), and \(p_v=p_u\,p_{u^\perp}\), then the full weighted theta kernel factors through a relative theta function \(\Theta_{L,M}\), and the corresponding theta contraction produces the restriction of a weighted theta lift from \(L\) to a sub-Grassmannian attached to \(M\) [2009.06012]. In this setting, polynomial weighting is built directly into the decomposition and contraction mechanism.

In the orthogonal Grassmannian setting, the polynomial insertion becomes even more explicit. For a quadratic space \(V\) of signature \((2,b_-)\) and \(Z\in G(V)\), the basic polynomial factors are
\[
P_{r,s,t}(\lambda,Z)
=
\frac{(\lambda,Z_{V,Z})^r(\lambda,\overline{Z_{V,Z}})^t}{(Y^2)^s},
\qquad
P_{r,s,t}^{(l)}(\lambda,Z)
=
P_{r,s,t}(\lambda,Z)(\lambda^2)^l,
\]
and the theta kernel summand is
\[
F_{r,s,t}(\tau,Z,\lambda)
=
e^{-\Delta_v/8\pi y}\bigl(P_{r,s,t}^{(l)}\bigr)(\lambda,Z)\,
e\!\left(\tau\frac{\lambda_+^2}{2}+\overline{\tau}\frac{\lambda_-^2}{2}\right)
\]
[1403.3567]. The resulting \(\Theta_{L,r,s,t}(\tau,Z)\) is automorphic of weight \((s-r,s-t)\) in \(Z\) and of weight \(\left(1+r+t,\ 2l+\frac{b_-}{2}\right)\) in \(\tau\) [1403.3567]. The same paper then studies how Grassmannian weight-raising and weight-lowering operators act on these polynomially weighted kernels and on their theta lifts.

## 4. Coding-theoretic and combinatorial realizations

A distinct realization of polynomially weighted theta functions appears in coding theory. For a binary linear code \(C\subset F_2^n\) and a discrete harmonic polynomial \(Q\), the harmonic weight enumerator is
\[
W_{C,Q}(x,y)=\sum_{c\in C}Q(c)\,x^{n-\mathrm{wt}(c)}y^{\mathrm{wt}(c)},
\]
which is presented as the precise coding-theoretic analogue of the weighted lattice theta series \(\Theta_{L,P}\) [1111.2392]. The analogy is structural: Euclidean norm shells correspond to Hamming weight shells, harmonic polynomials correspond to discrete harmonic polynomials, and the Poisson-summation transformation law corresponds to the generalized MacWilliams identity
\[
W_{C,Q}(x,y)
=
\left(-\frac{xy}{x^2-y^2}\right)^d
\frac{2^{n/2+d}}{|C^\perp|}
W_{C^\perp,Q}\!\left(\frac{x+y}{\sqrt2},\frac{x-y}{\sqrt2}\right)
\]
for \(Q\in \mathscr D_d^0\) [1111.2392].

For codes over \(\mathcal O_K/p\mathcal O_K\) attached to imaginary quadratic fields, the theta series of the Construction A lattice is obtained by evaluating a polynomial weight enumerator on coset theta functions. If \(C\) is a code and \(\mathrm{cwe}_C\) its complete weight enumerator, then
\[
\theta_{\Lambda_\ell(C)}(q)
=
\mathrm{cwe}_C\bigl(
\theta_{\Lambda_{0,0}}(q),
\theta_{\Lambda_{1,0}}(q),
\theta_{\Lambda_{2,0}}(q),
\dots,
\theta_{\Lambda_{p-1,p-1}}(q)
\bigr),
\]
and after identifying equal coset theta series this passes to the symmetric weight enumerator \(\mathrm{swe}_C\) [1309.3812]. In this framework the polynomial does not weight a lattice summand directly; rather, it controls how a finite family of theta functions is assembled into a new theta series. The same paper shows that different symmetric weight enumerators can produce the same theta function at a fixed admissible level, so the substitution map from polynomial data to theta series need not be injective [1309.3812].

A related bridge begins with matroids and Tutte polynomials. For a binary code \(C\), the standard identity
\[
w_C(\theta_3,\theta_2)=\theta_{L_C}(q)
\]
expresses the Construction A lattice theta series as a polynomial substitution into the code weight enumerator [2007.15089]. Combined with Greene’s theorem, this places the theta series under indirect control of the Tutte polynomial. The same paper treats this as the closest relevant analogue to “polynomial weighted theta functions” in its own framework [2007.15089].

## 5. Integrable-systems usage: finite-gap theta functions and quadratic-exponential weights

In the KdV literature, the nearest exact counterpart to a polynomially weighted theta function is not a theta function multiplied by a polynomial, but a finite-gap theta function multiplied by the exponential of a quadratic polynomial in the KdV times. For a hyperelliptic finite-gap solution, the tau-function is
\[
\tau(\mathbf t)
=
e^{\frac12\sum_{i,j\ge0}q_{ij}t_it_j}\,
\theta\!\left(\sum t_k\mathbf V^{(k)}-\mathbf u_0\right),
\]
so
\[
\log \tau(\mathbf t)
=
\frac12\sum_{i,j\ge0}q_{ij}t_it_j
+
\log\theta\!\left(\sum t_k\mathbf V^{(k)}-\mathbf u_0\right)
\]
[1807.03377]. The paper explicitly states that for finite-gap solutions the tau-function “coincides with the hyperelliptic theta-function up to multiplication by exponential of a quadratic polynomial” [1807.03377].

The main approximation theorem of the same work says that for any formal KdV tau-function \(\tau(\mathbf t)\) and any truncation order \(m>0\), there exists a hyperelliptic curve \(C\) of genus at most \(\lfloor m/2\rfloor\) and \(\mathbf u_0\in J(C)\setminus \Theta\) such that
\[
[\log\tau(\mathbf t)]_m
=
\left[
\log\theta\!\left(\sum_{2i+1\le m}t_i\mathbf V^{(i)}-\mathbf u_0\right)
\right]_m
+
\alpha
+
\sum_{2i+1\le m}\beta_i t_i
+
\frac12\sum_{2(i+j+1)\le m}\gamma_{ij}t_it_j
\]
[1807.03377]. Thus the non-quadratic part of \(\log\tau\) is reproduced by the logarithmic expansion of a finite-genus theta function, while the discrepancy is at most quadratic in the times.

This usage is conceptually close to polynomial weighting but technically different. The extra factor is \(e^{Q(\mathbf t)}\) with \(Q\) quadratic, and the comparison is carried out at the level of \(\log\tau\). The paper’s Witten–Kontsevich example makes this concrete by showing that the \(9\)-truncation of \(\log\tau\) coincides, modulo linear and quadratic terms, with the \(9\)-truncation of the logarithm of a genus \(4\) theta function built from the corresponding spectral curve [1807.03377].

## 6. Scattering, quantum, and operator-algebraic generalizations

Scattering-diagram theta functions provide a different nonclassical meaning of weighting. For a consistent scattering diagram \(\mathfrak D\), the theta function attached to \(p\) and endpoint \(Q\) is
\[
\vartheta_{p,Q}
=
\sum_{\operatorname{Ends}(\gamma)=(p,Q)} a_\gamma,
\]
a sum over broken lines [1503.06183]. The coefficients in products of theta functions are then expressed as weighted tropical counts:
\[
\alpha_Q(p_1,\dots,p_s;p)
=
\sum_{r\in K\cap P}
\left(
\sum_{\mathbf w\in \mathcal W_{\mathbf p}(\varphi(p)+r)}
\frac{N^{\mathrm{trop}}_{\mathbf w,\mathbf p}(Q)}{|\operatorname{Aut}(\mathbf w)|}
\right)
\]
[1503.06183]. In the quantum torus case, the weights are refined \(q\)-multiplicities such as
\[
[a]_q=q^a-q^{-a},
\]
and the paper states that these quantum multiplicities extend Block–Götsche multiplicities [1503.06183]. Here the “weighted theta function” is a theta function whose coefficients are tropical counts with Laurent-polynomial quantum weights.

A graph-theoretic generalization appears for non-commutative graphs. If \(S\subseteq \mathcal L(A)\) is a non-commutative graph and \(W\in\mathcal L(A)_+\), the weighted theta quantity is
\[
\widetilde{\vartheta}(S,W)
=
\min\left\{
\lambda:
Y\in S\otimes \mathcal L(B),\
\operatorname{Tr}_A Y\le \lambda I,\
Y\ge |\sqrt W\rangle\langle \sqrt W|
\right\},
\]
with \( |\sqrt W\rangle=(\sqrt W\otimes I)|\Phi\rangle \) [2101.00162]. The classical theory is recovered when \(W\) is diagonal. This is a weighted-theta formalism, but the weight is a positive semidefinite matrix rather than a polynomial.

A further nonclassical use occurs in quantum loop algebras and Yangians. There the “Theta series” are algebra-valued formal series whose weight components are indexed by the positive root cone \(Q^+\). If \(p\) has coweight \(\zeta\), then for every \(\beta\in Q^+\),
\[
\Theta_{p,\beta}(z),\ \overline{\Theta}_{p,\beta}(z)
\]
are polynomials in \(z\) with coefficients in
\[
U_q^-(\mathfrak g)_{-\beta}\otimes U_q^-(\mathfrak g)_\beta
\]
of degree bounded by \((\zeta,\beta)\) [2307.04043]. In this setting “weighted” refers to root-lattice grading, and polynomiality refers to the spectral parameter.

## 7. Analytic kernels, derivative generation, and theta-like series

The Kronecker theta function supplies a classical analytic mechanism for generating polynomial weights. It is defined by
\[
K_y(z|\tau)
=
\frac{\theta_1'(0|\tau)\,\theta_1(z+y|\tau)}
{\theta_1(z|\tau)\theta_1(y|\tau)},
\]
and Kronecker’s identity gives its bilateral expansion
\[
\sum_{k=-\infty}^{\infty}
\frac{e^{2kiy}}{1-q^k e^{2iz}}
=
\frac{i\,\theta_1'(0|\tau)\theta_1(y+z|\tau)}
{2\,\theta_1(y|\tau)\theta_1(z|\tau)}
\]
[2012.01670]. Liu’s decomposition theorem then states that a meromorphic function \(f(z)\) with the prescribed quasi-periodicity and only simple poles can be written as a finite sum of shifted Kronecker kernels:
\[
f(z)=\sum_{k=1}^n \operatorname{res}(f;a_k)\,K_y(z-a_k|\tau)
\]
[2012.01670]. The paper itself does not formulate a general theory of polynomially weighted theta sums, but this kernel makes such weights accessible by differentiation in the auxiliary parameter \(y\), since derivatives of \(e^{2kiy}\) produce powers of \(k\).

Theta-like functions also appear in asymptotic spline theory. The functions
\[
\varTheta_0(t)=1-\frac{4}{\pi}\sum_{k=0}^{\infty}(-1)^k\frac{e^{-(\pi^2/4)(2k+1)^2t^{-1}}}{2k+1},
\qquad
\varTheta_1(t)=1+2\sum_{k=1}^{\infty}(-1)^k e^{-(\pi k)^2t^{-1}}
\]
arise as limiting Mellin kernels for scaled \(B\)-splines, and the auxiliary series
\[
\varTheta_0^*(x)
=
\sum_{k=0}^\infty (-1)^k(2k+1)\exp\!\left(-(\pi^2/4)(2k+1)^2x\right)
\]
is an explicit theta-like series weighted by the polynomial \(2k+1\) in the summation index [2601.06643]. This is one of the clearest literal appearances of a polynomially weighted theta-like sum in the cited literature.

Taken together, these constructions show that polynomial weighting may enter theta theory in at least four distinct ways: as a polynomial inserted into the lattice or Siegel summand, as a polynomial factor in an orthogonal theta kernel, as a polynomial invariant controlling theta-series coefficients through substitution, or as a polynomial dependence generated by differentiating an auxiliary theta kernel. The common theme is that theta objects remain the organizing analytic structure, while polynomial data controls their coefficients, transformations, or asymptotic behavior.

Source: https://www.emergentmind.com/topics/polynomial-weighted-theta-functions