---
title: Symplectic Zauner's Conjecture
url: https://www.emergentmind.com/topics/symplectic-zauner-s-conjecture
type: topic
---

# Symplectic Zauner's Conjecture

Symplectic Zauner’s Conjecture most commonly denotes the strengthened Weyl–Heisenberg formulation of Zauner’s SIC conjecture: for every finite dimension \(d\), there should exist a Weyl–Heisenberg covariant SIC fiducial \(|\psi\rangle \in \mathbb C^d\) such that \(U(F_Z)|\psi\rangle = e^{i\theta}|\psi\rangle\) for a canonical order-3 symplectic matrix \(F_Z = \begin{pmatrix}0&-1\\1&-1\end{pmatrix}\). In this form, the conjecture is stronger than mere SIC existence, because it requires a specific symplectic symmetry in addition to the SIC overlap equations. A distinct recent literature uses the same name for a real symplectic equiangular tight frame existence problem, where equiangularity is measured by a skew-symmetric bilinear form and the maximal sizes are \(n=d\) or \(n=d+1\) rather than \(d^2\) [1908.02801] [2501.03970] [2509.14463].

## 1. Classical SIC and Weyl–Heisenberg formulations

Zauner’s original conjecture asserts that for each integer \(d \ge 2\) there exists a symmetric informationally complete POVM, equivalently an equiangular tight frame of \(N=d^2\) unit vectors in \(\mathbb C^d\). Concretely, one seeks unit vectors \(\{\psi_j\}\subset \mathbb C^d\) satisfying
\[
\sum_{j=1}^{d^2} \psi_j \psi_j^\dagger = \frac{d^2}{d} I_d,
\qquad
|\langle \psi_i,\psi_j\rangle|^2 = \frac{1}{d+1}
\quad (i\neq j),
\]
or, for the rank-1 projectors \(\Pi_j = |\psi_j\rangle\langle \psi_j|\),
\[
\operatorname{tr}(\Pi_i\Pi_j)=\frac{1}{d+1}, \quad i\neq j.
\]
In frame-theoretic language this is the \(N=d^2\) case of an ETF, where the coherence is \(c=1/\sqrt{d+1}\) [1908.02801].

The Weyl–Heisenberg covariant form replaces an arbitrary \(d^2\)-tuple by the orbit of a single fiducial under discrete displacement operators. In the standard computational basis,
\[
X|j\rangle = |j+1\rangle,\qquad Z|j\rangle = \omega^j |j\rangle,\qquad \omega=e^{2\pi i/d},
\]
and with \(\tau=-e^{\pi i/d}\) one defines
\[
D_{(p,q)}=\tau^{pq}X^pZ^q.
\]
A WH-covariant SIC is then characterized by a fiducial \(|\psi\rangle\) for which
\[
|\langle \psi|D_{(p,q)}|\psi\rangle|^2=\frac{d\,\delta_{(p,q),(0,0)}+1}{d+1}
\]
for all \((p,q)\in \mathbb Z_d^2\), together with the tight-frame identity
\[
\sum_{(p,q)\in\mathbb Z_d^2} D_{(p,q)} \Pi D_{(p,q)}^\dagger = d\,I
\]
in the rank-1 case [2501.03970].

The symplectic strengthening adds the requirement that the fiducial can be chosen in an eigenspace of a canonical order-3 Clifford or metaplectic unitary. Zauner observed numerically that a WH-covariant SIC fiducial can be chosen to satisfy
\[
U(F_Z)|\psi\rangle=e^{i\theta}|\psi\rangle,
\qquad
F_Z=\begin{pmatrix}0&-1\\1&-1\end{pmatrix},
\]
and this symmetry “drastically reduces the effective search space for fiducials and appears in all known exact and numerical solutions” [1908.02801]. In the extended Clifford formulation, “canonical order 3” means trace \(d-1\) modulo \(\bar d\), where \(\bar d=d\) for odd \(d\) and \(\bar d=2d\) for even \(d\) [2501.03970].

## 2. Symplectic and Clifford structure

The Weyl–Heisenberg operators form a projective representation of \(\mathbb Z_d^2\), and the symplectic group acts on phase-space indices. For \(F\in \operatorname{Sp}(2,\mathbb Z_d)\), the metaplectic unitary \(U(F)\) satisfies
\[
U(F)D_{p,q}U(F)^\dagger = D_{(p,q)F},
\]
while in the extended Clifford-group notation one writes
\[
U_F D_p U_F^\dagger = D_{Fp}.
\]
This is the structural origin of the adjective “symplectic” in the conjecture: the order-3 symmetry is imposed at the level of the phase-space action, not merely at the level of a unitary eigenspace [1908.02801] [2501.03970].

A standard form of the same phenomenon appears in the Appleby convention for the Clifford action,
\[
U_F D_p U_F^\dagger = e^{i\phi(F,p)}D_{Fp},
\]
and for the Zauner unitary one can write
\[
U_Z D_p U_Z^\dagger = \tau^{(\epsilon,p)} D_{F_Z p},
\qquad
U_Z|\psi\rangle=e^{i\theta}|\psi\rangle
\]
for some \(\epsilon\in \mathbb Z_d^2\). The literature cited here emphasizes the “so-far unexplained fact” that every known WH SIC fiducial is an eigenvector of an order-3 unitary of this type, and exploits that symmetry in both structural and constructive arguments [1903.06721].

This symplectic formulation is stronger than the weak existence statement for SICs. A SIC may exist without any specified covariance or eigenspace condition; Symplectic Zauner’s Conjecture requires a WH orbit and a canonical order-3 symmetry simultaneously. A plausible implication is that the conjecture is not only an existence problem for equiangular line sets, but also a rigidity statement about the organizing symmetries of fiducials.

## 3. Biangular Gabor frames and the topological route

A non-constructive route to Zauner’s conjecture replaces equiangular WH orbits by a larger class of biangular Gabor frames and then seeks to recover equiangularity by continuity [1908.02801]. Writing \(T\) for translation and \(M\) for modulation,
\[
(Tv)(j)=v(j-1),\qquad (Mv)(j)=\omega^j v(j),
\]
the WH orbit of \(v\in \mathbb C^d\) is
\[
G(v)=\{M^\ell T^k v\}_{k,\ell=0}^{d-1}.
\]
The orbit is \((\alpha,\beta)\)-biangular if
\[
|\langle v,T^k v\rangle|^2=\alpha \quad (k=1,\dots,d-1),
\]
and
\[
|\langle v,M^\ell T^k v\rangle|^2=\beta
\]
for every \(k\in\{0,\dots,d-1\}\) and \(\ell\in\{1,\dots,d-1\}\). Thus the nontrivial displacements split into the pure translation line and the complementary set.

The fundamental structural identity is the angle-balance lemma:
\[
\alpha + d\beta = \|v\|_2^4.
\]
For \(\|v\|_2=1\), this becomes \(\alpha+d\beta=1\). The WH orbit is always a \(d\)-tight frame, and the fourth-moment frame potential
\[
FP_4(v)=\frac{1}{d}\sum_{k,\ell=0}^{d-1} |\langle v,M^\ell T^k v\rangle|^4
\]
obeys
\[
FP_4(v)\ge \frac{2}{d+1},
\]
with equality if and only if \(v\) is a SIC fiducial. Under the biangular hypothesis,
\[
FP_4(\alpha,\beta)=\frac{1}{d}\Big[1+(d-1)\alpha^2+(d^2-d)\beta^2\Big],
\]
and using \(\beta=(1-\alpha)/d\) gives a quadratic minimized at
\[
\alpha=\beta=\frac{1}{d+1}.
\]
Among unit-norm biangular WH-covariant frames, the equiangular configuration is therefore the unique minimum of the WH 2-design frame potential.

The strategy is to study the real-algebraic variety \(B_d\subset \mathbb C^d\) of seeds whose WH orbits are biangular. After quotienting by nonzero scalars and fixing a coordinate, one obtains a slice
\[
C_d=\{v\in B_d: v(0)=1\}.
\]
Empirically, \(B_d/\mathbb C^\times\) and \(C_d\) “frequently appear to be path-connected and have low effective dimension despite being defined by \(\Omega(d^2)\) polynomial constraints.” The main continuity lemma states: if a Gabor MUB exists in \(\mathbb C^d\), meaning a seed with \((\alpha,\beta)=(0,1/d)\), and if \(B_d\) is path-connected, then a SIC exists. The argument connects that MUB-like seed to the normalized all-ones vector, which has \((\alpha,\beta)=(1,0)\), and uses the intermediate value theorem on \(\beta(t)-\alpha(t)\) along a path to force \(\alpha=\beta\).

The same paper proposes a symplectic refinement. Define the Zauner-invariant slice
\[
B_d^{Z,\theta}=\{v\in B_d: U(F_Z)v=e^{i\theta}v\}.
\]
If one can show nonemptiness and path-connectivity of \(B_d^{Z,\theta}\), and find two unit-norm biangular seeds on opposite sides of the SIC angle \(1/(d+1)\), then the same intermediate value argument yields a Zauner-symmetric SIC fiducial. The paper does not prove these steps, but formulates them as the geometric and topological core of a possible unconditional non-constructive proof. It gives an exact \(d=2\) analysis, where \(C_2\) is the union of two intersecting circles
\[
x^2+(y\pm1)^2=2
\]
for \(v=(1,x+iy)\), and numerical evidence in \(d=4\) and \(d=5\) that long continuous trajectories pass from MUB-like to trivial-like configurations and cross \(\alpha=\beta\) [1908.02801].

## 4. Conditional constructive resolution via Stark conjectures

A sharply different approach is constructive and arithmetic. The paper “A Constructive Approach to Zauner’s Conjecture via the Stark Conjectures” develops a conditional construction of WH-covariant SICs in all \(d>3\), together with a precise symplectic formulation of the conjecture, by using the order-1 abelian Stark conjecture for real quadratic fields and a special-value identity for the Shintani–Faddeev modular cocycle [2501.03970].

The basic object is a “ghost SIC,” with fiducial projector
\[
\tilde{\Pi}=\frac{r}{d}I+\sqrt{\frac{r(d-r)}{d^2(d^2-1)}}\sum_{p\not\equiv 0\ (d)} \tilde{\nu}_{Gp}\,D_p,
\]
where the normalized ghost overlaps are
\[
\tilde{\nu}_p=\phi_t(p)\,\shin_{A_t}^{d^{-1}p}(\rho_t),
\qquad
\tilde{\nu}_p\tilde{\nu}_{-p}=1.
\]
Here \(\shin_A^r(\tau)\) is the Shintani–Faddeev modular cocycle, a meromorphic function satisfying the multiplicative cocycle law
\[
\shin_{AB}^r(\tau)=\shin_A^r(B\cdot\tau)\,\shin_B^r(\tau),
\]
and at real quadratic fixed points \(\rho\),
\[
\shin_A^r(\rho)\,\shin_A^{-r}(\rho)=\psi^2(A)\,\chi_r(A).
\]

Two conjectural inputs drive the construction. The first is the Stark–Tate form of the order-1 abelian Stark conjecture for real quadratic fields, used to place the relevant special values in abelian extensions and to control their square roots. The second is the “Twisted Convolution Conjecture,” a special-value identity of the form
\[
\sum_{q \in \mathbb I_p}
\tau^{r\langle p,(\lambda I + Z)q\rangle}\,
\shin_{A_t^{-1}}^{d^{-1}q}(\rho_t)\,
\shin_{A_t^{-1}}^{d^{-1}(q-p)}(\rho_t)
=
d^2\,\delta_{p,0}^{(d)},
\]
which is precisely the identity needed to prove \(\tilde{\Pi}^2=\tilde{\Pi}\). Under these conjectures, the paper proves that the Galois conjugate of the ghost fiducial is a live SIC or, more generally, a live \(r\)-SIC fiducial, with nontrivial displacement overlaps of modulus
\[
|{\rm Tr}(\Pi D_p^\dagger)|=\sqrt{\frac{r(d-r)}{d^2-1}}
\quad (p\neq 0).
\]

The symplectic content is explicit. The construction works inside the extended Clifford group \(EC(d)\), where \(U_F D_p U_F^\dagger = D_{Fp}\). Associated stabilizers \(A_t\in \Gamma(d)\) and Zauner generators \(Z_t\) satisfy \(Z_t^{\,2m+1}=A_t\), and \(U_{Z_t}\) is a canonical order-3 unitary imposing Zauner symmetry on the fiducial. The paper therefore presents a conditional constructive resolution of the symplectic form of Zauner’s Conjecture.

Computationally, the construction is cross-validated against known exact WH SIC solutions and the Scott–Grassl catalogue. In \(d=100\), it produces four numerical examples of nonequivalent SICs, three of which are new. The paper also extends the framework to \(r\)-SICs for all \(r,d\) such that \(r(d-r)\) divides \((d^2-1)\), and studies the associated abelian field extensions over \(K=\mathbb Q(\sqrt{\Delta})\) with \(\Delta=(d+1)(d-3)\) in the rank-1 case [2501.03970].

## 5. Relaxations, associated structures, and continuous manifolds

A third line of work studies structures naturally attached to SICs and uses them to formulate relaxations of the WH SIC problem [1903.06721]. If \(G\) is the \(d^2\times d^2\) Gram matrix of a SIC and \(G^{(2)}=G\circ G\) is the Hadamard square, then
\[
Q=\frac{d+1}{2d}\,G^{(2)}
\]
is a projector of rank \(d(d+1)/2\) in dimension \(d^2\). From this one obtains
\[
H=2Q-I,
\]
a complex Hermitian Hadamard matrix of order \(d^2\), and two Naimark-complementary ETFs with Gram matrices
\[
E=\frac{2d}{d+1}\,Q,
\qquad
\bar E=\frac{d}{d-1}\left(\frac{I}{2}-Q\right),
\]
living in dimensions \(d(d+1)/2\) and \(d(d-1)/2\), respectively.

For odd \(d\), a WH SIC also yields two WH-covariant symmetric tight fusion frames of ranks \((d\pm1)/2\). Writing
\[
e^{i\theta_p}=\frac{\sqrt{d+1}}{d}\,{\rm Tr}(D_p\Pi),
\]
and
\[
A=\frac{1}{d}\sum_{p\in \mathbb Z_d^2} e^{2i\theta_{Fp}}D_p,
\qquad
\Pi_\pm=\frac{1}{2}(I_d\pm A),
\]
the resulting families \(\{D_p\Pi_\pm D_p^\dagger\}\) are STFFs with maximal number \(d^2\) of fusion elements. The core identity is the phase-squared convolution equation
\[
\sum_{u\in\mathbb Z_d^2}\tau^{-2(u,p)}e^{2i(\theta_u+\theta_{p-u})}=0
\quad (p\neq 0),
\]
which is considerably simpler than the full SIC equations.

This motivates two relaxations. The first asks only for phases \(\{e^{i\phi_p}\}\) with \(\phi_0=0\), \(\phi_{-p}=-\phi_p\), satisfying
\[
\sum_{u\in\mathbb Z_d^2}\tau^{-2(u,p)}e^{i(\phi_u+\phi_{p-u})}
=
d^2\,\delta_{p,0}
\]
for all \(p\in\mathbb Z_d^2\). For odd \(d\), any solution yields WH-covariant STFFs. The second uses the Naimark complement and a block-diagonal representation \(\widetilde D_p\) acting on \(\mathbb C^{d(d-1)}\), with relaxed constraints
\[
|\langle v|\widetilde D_p|v\rangle|^2
=
\frac{1}{(d-1)(d^2-1)}
\quad (p\neq 0).
\]
In this formulation the number of real variables scales as \(d^2\), while the number of equations is \(d^2-1\), so the equation-to-variable ratio tends to \(1\).

The same paper gives evidence that these associated structures lie on continuous manifolds even when the SIC fiducials themselves appear isolated. Explicit one-parameter families are constructed in \(d=3\), \(d=4\), \(d=6\), and \(d=8\), with affine families in \(d=3,4\) and non-affine families in \(d=6,8\). Restricted defect calculations for the structures arising from known SICs in \(d=2,\dots,16\) are nonzero for all \(d\ge 3\), with \(d=2\) the only isolated case. This suggests that the “phase-squared layer” around SICs is much more flexible than the fiducial layer itself, and that order-3 symplectic symmetry may be easier to study in these relaxed settings [1903.06721].

## 6. The real symplectic-space analogue

In a distinct usage, Fallon introduces equiangular tight frames in real symplectic spaces and formulates a “symplectic Zauner’s conjecture” for that setting [2509.14463]. Here the ambient space is a real symplectic space \(V=\mathbb R^{2d}\) with non-degenerate alternating form
\[
\omega(x,y)=x^TJy,
\qquad
J=
\begin{pmatrix}
0&I_d\\
-I_d&0
\end{pmatrix},
\]
or equivalently a canonically isomorphic model with block-diagonal matrix
\[
\Omega=\bigoplus_{i=1}^{d/2}\begin{bmatrix}0&1\\-1&0\end{bmatrix}.
\]
A frame \(\Phi=\{\phi_i\}_{i=1}^n\subset \mathbb R_{\mathcal S}^d\) is equiangular if there exists \(\mu>0\) such that
\[
|(\Phi^\dagger\Phi)_{ij}| = |[\phi_i,\phi_j]| = \mu
\quad (i\neq j),
\]
where the Gram matrix
\[
G=\Phi^\dagger\Phi
\]
is real skew-symmetric with zeros on the diagonal. Tightness is expressed not by the usual Euclidean identity, but by
\[
(\Phi\Phi^\dagger)^2=-c^2I_d
\qquad\Longleftrightarrow\qquad
(\Phi^\dagger\Phi)^3=-c^2(\Phi^\dagger\Phi),
\]
with \(\operatorname{rank}(G)=d\).

The real symplectic theory has a sharply different extremal regime. A symplectic Gerzon bound gives
\[
n\le d+1
\]
for equiangular sets, and the main existence theorem shows that a \(d\times n\) symplectic ETF can exist only if
\[
n=
\begin{cases}
d, & d\equiv 0 \pmod 4 \text{ or } d=2,\\
d+1, & d\equiv 2 \pmod 4.
\end{cases}
\]
Fallon’s conjecture is that these parameter values are also sufficient.

The main equivalence theorem states that this real symplectic conjecture is equivalent to the skew Hadamard conjecture. More precisely, for \(d>1\): a \(d\times d\) symplectic ETF exists if and only if there exists a skew Hadamard matrix of order \(d\), and a \(d\times(d+1)\) symplectic ETF exists if and only if there exists a skew Hadamard matrix of order \(d+2\). The proof translates the symplectic Gram matrix into the Seidel adjacency matrix of a tournament and analyzes the number of four-vertex subgraphs called diamonds. For odd \(n\),
\[
\delta_T\le \frac{1}{96}n(n-1)(n-3)(n+1),
\]
with equality characterizing, in the \(n\equiv 3\pmod 4\) case, switching equivalence to doubly regular tournaments. This combinatorial saturation supplies the flat kernel vector needed to complete a symplectic ETF Gram matrix to a skew Hadamard matrix.

The theory also includes explicit small examples, a doubling construction that lifts a skew Hadamard matrix of order \(d\) to one of order \(2d\), and a “complex-to-symplectic shadow” in which the imaginary part of the Gram matrix of certain complex ETFs yields, up to scaling, the Gram of a symplectic ETF. Although this is not the WH/Clifford conjecture of SIC theory, it is a mathematically precise and separate symplectic ETF existence theory that now shares the same name [2509.14463].

## 7. Status and unresolved directions

The current landscape is split between unconditional geometric programs, conditional arithmetic constructions, and relaxed frame-theoretic reformulations. In the biangular Gabor approach, the key open assumptions are path-connectivity of \(B_d\), existence of biangular seeds with \(\alpha<1/(d+1)\), and, for the symplectic version, nonemptiness and connectivity of the Zauner-invariant slice \(B_d^{Z,\theta}\). The paper explicitly formulates the connectivity and construction problems and notes that an answer, or suitable variants restricted to Zauner eigenspaces, would yield an unconditional non-constructive proof via continuity [1908.02801].

In the Stark–cocycle approach, the central unresolved points are the order-1 abelian Stark conjecture and the Twisted Convolution Conjecture. The existence theorems show that these two conjectural inputs imply the symplectic form of Zauner’s Conjecture, but the construction remains conditional until both are proved in full generality. The same paper also isolates conjectural equalities between the fields generated by overlaps and explicit ray class fields, tying the SIC problem to a real-quadratic instance of Hilbert’s twelfth problem [2501.03970].

The relaxations built from squared phases, Naimark complements, and associated Hadamard or ETF structures do not by themselves prove SIC existence, but they expose algebraic systems with fewer constraints or more balanced variable counts and show that the order-3 symplectic symmetry is operationally useful in those enlarged spaces. The existence of continuous manifolds for the associated structures in several dimensions, together with nonzero restricted defect in all tested cases \(d\ge 3\), suggests that the rigid fiducial problem sits inside a substantially more flexible geometric envelope [1903.06721].

In the real symplectic-space theory, the decisive open problem is the skew Hadamard conjecture itself. Because the symplectic ETF existence pattern is equivalent to that conjecture, any progress on skew Hadamard matrices translates immediately to new symplectic ETFs, and conversely algorithmic search through symplectic frame potentials may provide new skew Hadamard matrices. The paper also points to improved diamond-count methods, new core constructions, and further development of the complex-to-symplectic shadow as natural next steps [2509.14463].

Taken together, these strands indicate that “Symplectic Zauner’s Conjecture” now names two related but non-identical programs. In the classical SIC literature it is the claim that WH-covariant SIC fiducials can always be chosen with canonical order-3 symplectic symmetry. In the newer real symplectic ETF literature it is an extremal existence conjecture governed by skew-symmetric Gram matrices, tournaments, and skew Hadamard matrices. What unifies them is the central role of symplectic structure: in one case as Clifford covariance of SIC fiducials, in the other as the defining bilinear geometry of the frame space.

Source: https://www.emergentmind.com/topics/symplectic-zauner-s-conjecture