Persymmetric Matrix¶
A square matrix whose entries agree across reflection in its anti-diagonal, equivalently A = J Aᵀ J.
Core Idea¶
A persymmetric matrix is a square array that agrees with itself after reflection across its anti-diagonal. If J reverses row and column order, the test is A = J Aᵀ J. For an n×n matrix, entry (i,j) must equal entry (n−j+1,n−i+1). The transpose does not conjugate complex entries. This equation fixes which meaning of “persymmetric” is used here.[^ref-ec9002f3a32b]
The rule applies to many matrices, not one named example. Every Toeplitz matrix meets it. The population covariance of a real stationary time series is Toeplitz and therefore persymmetric; a mirror-symmetric Jacobi matrix used in spectral reconstruction also meets it.[ref-ec9002f3a32b][ref-08994455ed9a][^ref-659791b3e8dc]
Use Within the Domain¶
The same entry test appears in different matrix problems. In time-series statistics, a real stationary population covariance has entries determined by time lag. Inverse spectral theory studies real tridiagonal Jacobi matrices with mirrored coefficients. The covariance paper does not itself call its matrices persymmetric, and the spectral methods for the Jacobi subclass do not apply to every persymmetric matrix.[ref-08994455ed9a][ref-659791b3e8dc][^ref-ec9002f3a32b]
Clarity¶
Compare each entry with its anti-diagonal-reflected partner. Ordinary symmetry compares entries across the main diagonal; a Hankel matrix instead keeps values constant along each anti-diagonal. These are different tests. The exchange matrix J is the reversal operator used in the test, not another name for the whole class.[^ref-ec9002f3a32b]
An anti-diagonal matrix, whose off-axis entries are zero, always passes the persymmetry test. It is a narrower case: persymmetric matrices may have nonzero entries elsewhere. This is a consequence of the pairwise equality, not an alternative definition.[^ref-ec9002f3a32b]
Manages Complexity¶
The compact equation replaces many separate pairwise comparisons. It also supports one safe deduction: if a persymmetric matrix is invertible, its inverse remains persymmetric. Stronger structure need not survive; the inverse of a Toeplitz matrix, for example, need not remain Toeplitz.[^ref-ec9002f3a32b]
Abstract Reasoning¶
To refute persymmetry, one unequal reflected pair is enough. To establish it, check all pairs or prove the equivalent matrix equation. For a Toeplitz matrix, reflected positions have the same row-minus-column difference, which explains its inclusion. This reasoning is about the matrix entries, independent of what the matrix represents.[^ref-ec9002f3a32b]
Knowledge Transfer¶
The anti-diagonal test transfers from covariance matrices to Jacobi matrices, but application-specific conclusions do not. A statistical estimator's behavior does not follow from a Jacobi reconstruction theorem, and neither follows merely from persymmetry. The broader Matrix and Symmetry relations explain the shared array carrier and fixed-point structure without replacing this specialist equation.[ref-ec9002f3a32b][ref-08994455ed9a][^ref-659791b3e8dc]
Example¶
Xiao and Wu write the real stationary population covariance as Σ_T = (γₛ₋ₜ): its entries depend on the lag s−t. Reflection sends (s,t) to (T−t+1,T−s+1), preserving that lag, so the reflected entries agree. Mapped back: Σ_T is the square array, the index reversal supplies the pairing, and lag equality supplies the required values. This is a derived persymmetry example; their original paper names the Toeplitz property.[ref-08994455ed9a][ref-ec9002f3a32b]
Relationships to Other Abstractions¶
Current abstraction Persymmetric Matrix Domain-specific
Parents (2) — more general patterns this builds on
-
Persymmetric Matrix is a kind of Matrix Domain-specific
A persymmetric matrix is a Matrix with the additional equality A = J A^T J.
-
Persymmetric Matrix is a decomposition of Symmetry Prime
The matrix is fixed by the reversal-transpose involution, instantiating Symmetry.
Hierarchy paths (6) — routes to 6 parentless roots
- Persymmetric Matrix → Matrix → Tensor → Transformation → Function (Mapping)
- Persymmetric Matrix → Symmetry
- Persymmetric Matrix → Matrix → Linearity
- Persymmetric Matrix → Matrix → Representation → Abstraction
- Persymmetric Matrix → Matrix → Tensor → Invariance
- Persymmetric Matrix → Matrix → Tensor → Vector Space → Set and Membership
Neighborhood in Abstraction Space¶
Persymmetric Matrix sits in a sparse region of the domain-specific corpus (85th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Matrix Structures & Matroids (11 abstractions)
Nearest neighbors
- Matrix Similarity — 0.83
- Exchange matrix — 0.82
- Matrix Pencil — 0.81
- Anti-Diagonal Matrix — 0.80
- Roothaan–Hall Equations — 0.80
Computed from structural-signature embeddings · 2026-10-17
Not to Be Confused With¶
- Ordinary symmetric matrix: equality across the main diagonal, not the anti-diagonal.[^ref-ec9002f3a32b]
- Hankel matrix: constant values along anti-diagonals; the rule differs from reflection across one.[^ref-ec9002f3a32b]
- Anti-diagonal matrix: a narrower sufficient case, not the whole class.[^ref-ec9002f3a32b]
- Complex conjugate-reversal condition: may coincide under additional assumptions, but the definition here uses transpose without conjugation.[^ref-ec9002f3a32b]
References¶
[^ref-ec9002f3a32b]: Peter Kabal, Minimum Mean-Square Error Filtering, Autocorrelation/Covariance, General Delays, and Multirate Systems, version 0.985, McGill University, 2011. The printed report title uses a colon after “Filtering.” Full author-hosted technical monograph inspected. Appendix C.4, printed pp. 166–167, equations C.8–C.12, defines transpose-only persymmetry and proves inverse closure; C.5–C.6, pp. 167–168, distinguishes Toeplitz and Hankel structures. This is technical exposition, not the original discovery of the matrix class.
[^ref-08994455ed9a]: Han Xiao and Wei Biao Wu, “Covariance Matrix Estimation for Stationary Time Series”, Annals of Statistics 40 (2012), 466–493, DOI 10.1214/11-AOS967. Full original paper inspected, Introduction, equations (2)–(3) and adjacent discussion. It establishes the real stationary covariance's Toeplitz form; the persymmetric inclusion follows separately from Kabal's relation.
[^ref-659791b3e8dc]: Vincent X. Genest, Satoshi Tsujimoto, Luc Vinet, and Alexei Zhedanov, “Persymmetric Jacobi Matrices, Isospectral Deformations and Orthogonal Polynomials”, arXiv:1605.00708v2, 2017. Full original paper inspected, Abstract and §§1–2 for the mirror-symmetric Jacobi identity and §§3–4 for its restricted inverse spectral and deformation analysis.