---
title: Peres Test in Quantum Diagnostics
url: https://www.emergentmind.com/topics/peres-test
type: topic
---

# Peres Test in Quantum Diagnostics

In the literature surveyed here, the term **Peres test** labels several distinct constructions associated with Asher Peres. The best-known usage is the **state-independent** test of **noncontextual realist models** based on the Peres–Mermin square for a **two-qubit system** [1005.2402]. A second usage denotes Peres’s interferometric proposal for probing whether phase shifts behave as ordinary **complex** quantities or as **hyper-complex** ones such as quaternions [1604.04950]. A third usage appears in quantum chaos, where **Peres lattices** visualize regularity and chaos by plotting expectation values of observables across eigenstates or Floquet modes [1312.2555]. Across these settings, the common role of the test is diagnostic: a compact algebraic, interferometric, or spectral structure is used to separate quantum predictions from an alternative description.

## 1. Terminological scope and principal senses

The expression **Peres test** is not confined to one protocol. In the sources considered here, it refers to several technically distinct constructions.

| Usage | Core object | Diagnostic target |
|---|---|---|
| Peres–Mermin test | \(3\times 3\) array of commuting two-qubit observables | Noncontextual realist models |
| Peres interferometric test | Order of two phase-shifting devices in an interferometer | Hyper-complex or quaternionic phase noncommutativity |
| Peres test on quantum hardware | Quantity \(F\) built from pairwise overlaps in a three-level superposition | Whether amplitudes behave as complex numbers |
| Peres lattices | Scatter plots of expectation values across eigenstates or Floquet modes | Regularity, resonances, and chaos |

A related but not identical vocabulary also appears in later work. The literature uses **Peres-type criterion** for steering tests based on the spectrum of the partial transpose [2601.12085], **Peres conjecture** for contextuality minimality [2001.07656], and **Peres conjecture** for Bell nonlocality versus distillability [1405.4502]. These developments are part of the broader Peres lineage, but they are not the same construction as the Peres–Mermin square or the interferometric phase-order test.

## 2. The Peres–Mermin contextuality test

For a two-qubit system, the original Peres–Mermin argument uses nine Pauli-product observables arranged as

\[
\begin{array}{ccc}
\sigma_x^{(1)}\otimes I^{(2)} & I^{(1)}\otimes \sigma_x^{(2)} & \sigma_x^{(1)}\otimes \sigma_x^{(2)} \\
I^{(1)}\otimes \sigma_y^{(2)} & \sigma_y^{(1)}\otimes I^{(2)} & \sigma_y^{(1)}\otimes \sigma_y^{(2)} \\
\sigma_x^{(1)}\otimes \sigma_y^{(2)} & \sigma_y^{(1)}\otimes \sigma_x^{(2)} & \sigma_z^{(1)}\otimes \sigma_z^{(2)}
\end{array}.
\]

In **each row** and **each column**, the three observables are mutually commuting, so each row or column can be measured jointly. Quantum mechanically, the products of the three observables in the rows and columns satisfy

\[
R_1 = R_2 = R_3 = C_1 = C_2 = +I,\qquad C_3 = -I,
\]

equivalently,

\[
R_1|\psi\rangle=R_2|\psi\rangle=R_3|\psi\rangle=C_1|\psi\rangle=C_2|\psi\rangle=|\psi\rangle,
\]
\[
C_3|\psi\rangle=-|\psi\rangle
\]

for any state \(|\psi\rangle\) in the four-dimensional space. The contradiction with a noncontextual model is generated by the **product rule**

\[
v(AB)=v(A)v(B),
\]

with predetermined values \(v(A)=\pm1\) for commuting observables. Applying the rule to all six row and column products yields mutually inconsistent constraints: multiplying the hidden-variable assignments gives \(+1\), while the operator identities force an overall minus sign. The contradiction is therefore **state-independent** [1005.2402].

The experimentally testable form cited in this context is Cabello’s inequality,

\[
\langle \chi \rangle_{\mathrm{ncr}}=
\langle R_1\rangle+\langle R_2\rangle+\langle R_3\rangle+\langle C_1\rangle+\langle C_2\rangle-\langle C_3\rangle \le 4,
\]

whereas quantum mechanics predicts

\[
\langle \chi \rangle_{\mathrm{QM}} = 6
\]

for any state of the two-qubit system [1005.2402].

The 2010 variant replaces the four identity-containing observables by four new **holistic observables**, yielding a more symmetric \(3\times3\) array built entirely from two-qubit Pauli products. In that variant, the row and column products satisfy

\[
R'_1 = R'_2 = R'_3 = +I,\qquad C'_1 = C'_2 = C'_3 = -I,
\]

again for any quantum state, and the associated noncontextual inequality becomes

\[
\langle \chi \rangle_{\mathrm{ncr}}=
\langle R_1\rangle+\langle R_2\rangle+\langle R_3\rangle-\langle C_1\rangle-\langle C_2\rangle-\langle C_3\rangle \le 4,
\]

with the same quantum value \(\langle \chi \rangle_{\mathrm{QM}} = 6\) [1005.2402]. The conceptual change is not the logical structure of the contradiction, but the **choice of observables and the resulting inequality**.

## 3. Operationalization, robustness, and interpretive caveats

The algebraic contradiction by itself is not a complete experimental framework. A generalized treatment derives **noise-robust noncontextuality inequalities** by replacing ideal observables with operational equivalence classes of measurements and by introducing source procedures so that generalized noncontextuality constrains both measurements and preparations [1704.01153]. In that approach, the relevant experimentally accessible quantities are the nine source-measurement correlations

\[
\omega_{ij} \equiv \langle \mathcal{s}_{ij}\mathcal{m}_{ij}\rangle_{\mathcal{S}_{ij}\mathcal{M}_{ij}},
\]

and the set of noncontextual correlations forms a polytope with **184 facet inequalities**. This produces inequalities that are **necessary and sufficient** for the restricted nine-correlation scenario to admit a noncontextual model [1704.01153].

That same work criticizes earlier Cabello-style tests based directly on the six triple products. The reason is operational rather than algebraic: if the compatibility structure is realized, then the triple-product expectations already satisfy

\[
\langle m_{11}m_{12}m_{13}\rangle_P=
\cdots=
\langle m_{12}m_{22}m_{32}\rangle_P=+1,\qquad
\langle m_{13}m_{23}m_{33}\rangle_P=-1,
\]

so

\[
R(P)=6
\]

regardless of noise. On this analysis, the inequality \(R(P)\le4\) is not a robust test of generalized noncontextuality [1704.01153].

A different caveat concerns the physical realization of the square. It has been argued that the Peres–Mermin square does **not necessarily** rule out a **value-definite deterministic noncontextual hidden-variable model** unless two extra assumptions are satisfied: **(i) unique realization**, meaning each operator is represented by exactly one physical measurement, and **(ii) simultaneous measurability**, meaning commuting operators are represented by measurements that can be performed together in the same run [2011.06508]. Three explicit hidden-variable models were constructed for three physical realizations of the square: one violating (i), another violating (ii), and a third violating both. The point is not that quantum mechanics is noncontextual, but that the operator square becomes a contextuality proof only when the measurement realization matches the operator algebra in the required way [2011.06508].

A further operational question is the **classical memory cost** of simulating the Peres–Mermin scenario. For sequences of compatible measurements, a classical sequential automaton with exactly **three internal states** is sufficient to reproduce **all quantum sequential correlations** for **any quantum state**, and this is also optimal because three states are already necessary for the deterministic predictions [1611.07515]. In that precise sense, the Peres–Mermin scenario is contextual but not memory-expensive.

## 4. Peres interferometry and tests of complex versus quaternionic phases

In another established usage, the Peres test is an interferometric proposal concerning **hyper-complex** quantum theories. Peres’s idea is to send a beam through two materials \(A\) and \(B\) in opposite orders along two interferometer arms, compare the accumulated phases, and ask whether the order matters. In ordinary complex quantum theory the phases commute, so the order does not change observable outcomes; in a quaternionic theory, noncommuting phases could in principle produce a measurable difference [1604.04950].

A central analysis of this proposal considers standard quaternionic Hilbert-space quantum theory with a **spatially localized scattering interaction**. Under those assumptions, the Hamiltonian is written as

\[
H = H_0 + V,
\]

the \(S\)-matrix commutes with the free Hamiltonian,

\[
[H_0,S]=0,
\]

and on the energy shell one finds

\[
[i,S]=0.
\]

The conclusion is that the observable scattering matrix is **complex-valued**, not quaternionic. Because the refractive index is tied to the forward scattering amplitude, and that amplitude is built from the \(S\)-matrix, the refractive index relevant for the optical experiment is also **complex**, not quaternionic. On this analysis, the optical Peres-type experiment of Procopio et al. does **not** test quaternionic quantum mechanics within the standard Hilbert-space framework; genuine quaternionic effects would have to be sought in the **near-zone field**, not in the **radiation-zone field** [1604.04950].

A reply from a **black-box / instrumentalist generalized probabilistic theory** perspective makes a different claim. In that view, the experiment does not assume a microscopic quaternionic mechanism; it tests whether effective transformations commute in the way complex quantum theory predicts. The authors therefore maintain that the experiment places meaningful bounds on possible **post-quantum theories**, including quaternionic ones, and they exhibit a relativistic **Klein–Gordon scattering** example in which quaternionic components survive in transmission, with \(C_8\neq0\) outside the pure complex limit. They emphasize that this relativistic example lies outside Adler’s assumptions because the Klein–Gordon Hamiltonian formulation has an **indefinite metric** [1607.01648].

The same foundational question has been repurposed as a hardware benchmark. On a quantum computer, the Peres test is implemented by preparing a three-level superposition and constructing

\[
y_{ij} = 2\langle B(i)|\psi\rangle \langle \psi|B(j)\rangle,
\]

followed by

\[
F = |y_{01}| + |y_{12}| + |y_{20}| - 2|y_{01}y_{12}y_{20}|.
\]

The ideal prediction is

\[
F=1,
\]

which is interpreted as the signature that **complex numbers** are sufficient for quantum mechanics [2207.13585]. In this implementation, readout noise, depolarizing noise, and thermal relaxation noise affect the value of \(F\) differently. For the particular state studied, readout error shows a threshold behavior around \(p\approx0.9\), depolarizing noise drives \(F\) away from and then back toward \(1\), and thermal relaxation leads \(F\) toward its ideal value as \(T_1\) becomes much larger than all gate times [2207.13585].

## 5. Peres lattices as a spectral test of regularity and chaos

Peres lattices adapt the diagnostic idea to spectral analysis. In a stationary system, one chooses an operator \(P\) that is a constant of motion for some unperturbed Hamiltonian \(H_0\), computes \(\langle P\rangle\) in the eigenstates of the full Hamiltonian, and plots

\[
\langle P\rangle \quad \text{vs.} \quad E.
\]

If the system is integrable or nearly regular, the points form an ordered lattice; if chaos sets in, the lattice becomes irregular. In this sense the Peres lattice is a qualitative quantum analogue of a **Poincaré section** [1312.2555].

For the Dicke model,

\[
H=\omega a^{\dagger}a+\omega_{0}J_{z}+\frac{2\gamma}{\sqrt{\mathcal{N}}}(a+a^{\dagger})J_{x},
\]

the observables used were \(J_z\), \(J_x^2\), and \(n=a^\dagger a\). The analysis showed that regular and irregular regions can **coexist** in the spectrum, even **below** the superradiant critical coupling \(\gamma_c=\frac{\sqrt{\omega\omega_0}}{2}\) [1312.2555]. The \(J_z\) lattice also reveals two excited-state quantum phase transitions: a **static ESQPT** at \(E/j=1\), associated with saturation of the atomic subsystem, and a **dynamic ESQPT** at \(E/j=-1\), associated with the state \(n=0,\ m=-j\) and with the superradiant rearrangement of the spectrum [1312.2555].

The method has been extended from stationary systems to **periodically driven** ones by replacing stationary eigenstates with **Floquet modes**. If

\[
\widehat H(t)=\widehat H(t+T)=\widehat H_0+\eta\, g(t)\,\widehat H',
\]

the Floquet modes are defined by

\[
\widehat U(t_0+T,t_0)\,|F_i(t_0)\rangle = e^{-i F_i T}\,|F_i(t_0)\rangle,
\]

with quasienergies \(F_i\) taken in a Floquet-Brillouin zone such as

\[
F_i \in \left[-\frac{\pi}{T}, \frac{\pi}{T}\right).
\]

The driven-system lattice plots expectation values of two observables in the Floquet modes, because periodic driving increases the effective number of degrees of freedom by one [2606.20009]. In this form, Peres lattices efficiently detect resonances, monitor the onset of chaos, identify critical properties of Floquet modes, compare different driving Hamiltonians, and test iterative approximation techniques based on effective stationary Hamiltonians [2606.20009]. The lattices are therefore not merely visual aids; they provide a compact map of spectral organization, resonance structure, and ergodic spreading.

## 6. Related Peres-type descendants and conjectures

The broader Peres program generated several influential descendants. In four-dimensional Hilbert space, the **24 rays of Peres** form **24 orthonormal bases** and contain exactly **512 parity proofs** of the Kochen–Specker theorem, of types \(18\)-\(9\), \(20\)-\(11\), \(22\)-\(13\), and \(24\)-\(15\) [1103.6058]. These parity proofs sharpen the contextuality content of the Peres construction by reducing the contradiction to odd-versus-even counting of value assignments across bases.

The combinatorial minimization problem behind this program was settled when the **Peres conjecture for contextuality** was proved: the smallest possible Kochen–Specker vector set in any dimension has **18 vectors**, namely the Cabello–Estebaranz–García-Alcaine construction [2001.07656]. That result uses a graph-theoretic framework involving the **stable set polytope** \({\rm STAB}(G)\) for classical assignments and the **theta body** \({\rm TH}(G)\) for quantum assignments, and it shows that no GHZ-type proof exists with fewer than **10 events** [2001.07656].

A different conjecture of Peres concerned Bell nonlocality and distillability. It was disproved by constructing an explicit **two-qutrit PPT state** that is **bound entangled** and yet violates a Bell inequality. In that example, the state satisfies \(\mathrm{PT}(\varrho)=\varrho\), hence is undistillable, but still yields a strictly positive Bell value \(I_{\varrho}\simeq 2.63144\times 10^{-4}\) [1405.4502]. The conceptual consequence is that **Bell nonlocality implies neither entanglement distillability nor non-positivity under partial transposition** [1405.4502].

Finally, a recent **Peres-type criterion** has been proposed for **EPR steering** in two qubits. Writing the eigenvalues of \(\rho_{AB}^{T_B}\) as \(\{\lambda_1,\lambda_2,\lambda_3,\lambda_4\}\) and the corresponding elementary symmetric polynomials as \(\Lambda_2,\Lambda_3,\Lambda_4\), the steering functional is

\[
\mathcal{S} = (1+\mu)\Lambda_4\left(1-\frac{32}{9}\Lambda_2\right)+\Lambda_3^2,
\qquad \mu\in(0,2],
\]

with steerability detected by

\[
\mathcal{S}<0.
\]

This criterion is claimed to be **necessary and sufficient** for the two-qubit Werner state, for all two-qubit pure states, and for all two-qubit rank-2 states [2601.12085]. It extends the Peres strategy of using the partial transpose as a compact spectral diagnostic, but now for the intermediate nonlocality notion of steering rather than entanglement itself.

The term **Peres test** therefore designates a family of diagnostic constructions rather than a single object. Its central historical form is the Peres–Mermin state-independent test of noncontextuality, but the same label also covers phase-order interferometry for hyper-complex quantum theories, hardware tests of the complex-amplitude postulate, and lattice-based spectral diagnostics of regularity and chaos. The later Peres-type criteria and Peres conjectures show how this style of reasoning migrated into steering, Bell nonlocality, and the combinatorics of contextuality, while preserving the characteristic Peres strategy: isolate a small, rigid structure whose empirical realization sharply constrains any competing model.

Source: https://www.emergentmind.com/topics/peres-test