Three-State Geometric Tensors

• Three-state geometric tensors are tensorial structures defined by three independent directional states and symmetry constraints that govern their algebraic and geometric properties.
• They underpin diverse applications such as nematic liquid crystals, cross fields, and multipartite entanglement by bridging multilinear algebra with physical phenomena.
• Their framework informs spectral analysis, defect characterization, and entanglement measures, offering insights into condensed matter physics and algebraic statistics.

A three-state geometric tensor is a tensorial object or structure corresponding to a geometric "state space" with three independent directions, states, or components, governed by algebraic, topological, or physical constraints that single out special roles for the number three. Central examples span symmetric third-order tensors in three dimensions, linear-algebraic classifications based on geometric rank, geometric approaches to multi-band quantum observables, and tensorial models for nematic liquid crystals, cross fields, and multipartite entanglement. Such tensors underpin deep connections between multilinear algebra, invariant theory, topological defects, and multi-state quantum and statistical systems.

1. Algebraic and Geometric Foundations

Three-state geometric tensors are most classically realized as symmetric, traceless third-order tensors in three-dimensional space. The prototypical setting is the space

$$H(3,3) = \{ \mathcal{Q} \in \mathbb{R}^3 \otimes_s \mathbb{R}^3 \otimes_s \mathbb{R}^3 : \operatorname{tr}_{jk} \mathcal{Q} = 0 \}$$

where the trace is taken over any two indices. A canonical isomorphism exists between such tensors and the geometric data of a regular tetrahedron $\{u_1,\ldots,u_4\} \subset S^2$ via

$$T_{ijk} = \sum_{\ell=1}^4 u_i^\ell u_j^\ell u_k^\ell.$$

This identification encodes the four equiangular directions of a tetrahedral frame into tensorial language, with explicit constraints (such as $u_i \cdot u_j = -1/3 + (4/3)\delta_{ij}$). The nonlinear constraint

$\mathcal{Q} \mathcal{Q}^T = \frac{32}{27} I_3$

characterizes those tensors lying precisely on the SO(3)/T orbit associated with tetrahedral symmetry, reflecting a one-to-one correspondence up to rotation (Golovaty et al., 2022).

Parallel formulations occur for 3D cross fields, where invariant properties under the octahedral group $\mathbb{O} \subset SO(3)$ lead to a linear representation in terms of fully symmetric, trace-restricted fourth-order tensors living in a nine-dimensional real vector space. These frameworks generalize the classical Segre embedding and the algebraic varieties associated to symmetry groups.

2. Geometric Rank Three and Tensor Classification

The notion of geometric rank (GR) applies to tripartite tensors $T \in A \otimes B \otimes C$, with $$\mathrm{GR}(T) = \operatorname{codim}_{A^* \times B^*} \Sigma_T$$, where $\Sigma_T$ is the locus of “singular slices” in a fixed flattening. Three-state tensors are those with $\mathrm{GR}(T)=3$. According to (Geng, 2022), such tensors decompose into three mutually exclusive categories:

• Tensors whose any flattening has matrix rank at most $\{u_1,\ldots,u_4\} \subset S^2$0.
• Tensors of slice rank $\{u_1,\ldots,u_4\} \subset S^2$1.
• The primitive case, notably the $\{u_1,\ldots,u_4\} \subset S^2$2 matrix multiplication tensor $\{u_1,\ldots,u_4\} \subset S^2$3 and certain $\{u_1,\ldots,u_4\} \subset S^2$4 skew-symmetric matrix tensors.

Primitive geometric rank three tensors are maximally non-decomposable under this notion, and their ambient geometry links to linear determinantal varieties with controlled codimensions, connection to secant and join varieties, and explicit rank constraints for flattenings.

3. Topology, Symmetry, and Singular Structures

In variational and field-theoretical models, three-state geometric tensors enforce symmetry breaking, defect structures, and boundary conditions via their algebraic constraints. For instance, the tetrahedral frame field model with a Ginzburg–Landau relaxation

$\{u_1,\ldots,u_4\} \subset S^2$5

drives the tensor field to lie in the constrained orbit except along a singular set:

• In 2D, singularities are isolated vortices, with fractional index (e.g., $\{u_1,\ldots,u_4\} \subset S^2$6) and topological constraints on total index.
• In 3D, the singular set is a network of one-dimensional filaments meeting at triple junctions, classified by the nonabelian fundamental group $\{u_1,\ldots,u_4\} \subset S^2$7 (Golovaty et al., 2022).

These defect structures reflect the interplay between geometry (the target manifold), algebraic constraints (tracelessness and symmetry), and topology (homotopy classes of defects).

4. Three-State Quantum Geometric Tensors in Condensed Matter Physics

Quantum geometric tensors (QGTs) extend the notion of metric structure from differential geometry into the Bloch state bundle in multi-band quantum systems. For three bands ($\{u_1,\ldots,u_4\} \subset S^2$8), relevant quantities include:

• The interband quantum metric, $\{u_1,\ldots,u_4\} \subset S^2$9.
• The gauge-invariant shift vector, $$T_{ijk} = \sum_{\ell=1}^4 u_i^\ell u_j^\ell u_k^\ell.$$0.
• The triple-phase product (Bargmann invariant), $$T_{ijk} = \sum_{\ell=1}^4 u_i^\ell u_j^\ell u_k^\ell.$$1.

Three-state terms, such as $$T_{ijk} = \sum_{\ell=1}^4 u_i^\ell u_j^\ell u_k^\ell.$$2, generate injection and shift photocurrents under bicircular light even in centrosymmetric, time-reversal invariant systems, in contrast to the vanishing of traditional shift current effects in 2-band models (Guo et al., 5 Mar 2025).

5. Characteristic Polynomial, Eigenvalues, and Identifiability

Symmetric third-order tensors in three dimensions (plane cubic forms) admit a well-developed spectral theory. The characteristic polynomial mapping

$$T_{ijk} = \sum_{\ell=1}^4 u_i^\ell u_j^\ell u_k^\ell.$$3

is algebraic, finite-to-one, and all fibers (i.e., tensors with the same characteristic polynomial) are 0-dimensional in projective space (Galuppi et al., 2023). This implies finite identifiability up to isomorphism, with conjectural fiber size $$T_{ijk} = \sum_{\ell=1}^4 u_i^\ell u_j^\ell u_k^\ell.$$4 for this case. The eigen-equations formalize as

$$T_{ijk} = \sum_{\ell=1}^4 u_i^\ell u_j^\ell u_k^\ell.$$5

encoding the spectral structure of three-state symmetric tensors.

6. Entanglement Measures and Tensor Norms in Three-Partite Quantum Systems

Multipartite quantum entanglement, particularly in the tripartite ($$T_{ijk} = \sum_{\ell=1}^4 u_i^\ell u_j^\ell u_k^\ell.$$6) setting, admits a geometric/tensorial description. The geometric measure of entanglement (GME) for a pure state tensor $$T_{ijk} = \sum_{\ell=1}^4 u_i^\ell u_j^\ell u_k^\ell.$$7 is given by

$$T_{ijk} = \sum_{\ell=1}^4 u_i^\ell u_j^\ell u_k^\ell.$$8

where $$T_{ijk} = \sum_{\ell=1}^4 u_i^\ell u_j^\ell u_k^\ell.$$9 is the spectral norm (maximum overlap with a product state), and its dual, the nuclear norm $u_i \cdot u_j = -1/3 + (4/3)\delta_{ij}$0, corresponds to an “entanglement energy” $u_i \cdot u_j = -1/3 + (4/3)\delta_{ij}$1. For mixed states, separability is characterized by the nuclear norm of density tensors $u_i \cdot u_j = -1/3 + (4/3)\delta_{ij}$2 attaining unity. In the bosonic (fully symmetric) case, these criteria are polynomial-time computable for fixed dimensionality (Friedland, 25 Sep 2025). The unique maximally entangled three-qubit pure state (the W-state) achieves $u_i \cdot u_j = -1/3 + (4/3)\delta_{ij}$3, $u_i \cdot u_j = -1/3 + (4/3)\delta_{ij}$4, $u_i \cdot u_j = -1/3 + (4/3)\delta_{ij}$5.

7. Statistical Models and Algebraic Geometry of Three-State Tensors

In algebraic statistics, three-state tensors parameterize models of rater agreement for three raters (QI₃, qI₃, p–qI₃ and corresponding mixture models). These are rational projective varieties of dimensions $u_i \cdot u_j = -1/3 + (4/3)\delta_{ij}$6 to $u_i \cdot u_j = -1/3 + (4/3)\delta_{ij}$7, all of degree $u_i \cdot u_j = -1/3 + (4/3)\delta_{ij}$8, constructed via Segre embeddings, Hadamard products, or joins with diagonal loci. Their defining ideals are generated by minors of flattenings, higher-degree binomials, and specialized “tetra–nomials” for certain mixture models. For example, the quasi-independence model QI₃ is defined by all $u_i \cdot u_j = -1/3 + (4/3)\delta_{ij}$9 minors of the three flattenings, yielding $\mathcal{Q} \mathcal{Q}^T = \frac{32}{27} I_3$0 independent quadrics (Bocci et al., 2018). Pairwise Cohen’s $\mathcal{Q} \mathcal{Q}^T = \frac{32}{27} I_3$1 coefficients admit explicit rational parametrizations on these varieties, revealing the geometry of chance-corrected agreement.

---

Three-state geometric tensors thus function as unifying formal objects across multilinear algebra, geometry, topology, condensed matter physics, quantum information, and algebraic statistics, encoding both symmetrical structure and rich defect, spectral, and entanglement phenomena. They serve as a critical component in the algebraic, variational, and topological analysis of systems with inherent threefold structure, with applications from nematic liquid crystals and mesh generation to three-band quantum materials and multipartite entanglement.