Nonreciprocal Frustration: Asymmetric Dynamics

• Nonreciprocal frustration is a phenomenon where asymmetric interactions or directed couplings create incompatible dynamical objectives, replacing static degeneracies with continuous marginal orbits.
• It encompasses diverse models such as nonreciprocal Ising/XY systems, directed signed networks, and dissipative quantum spins, all of which illustrate complex spatiotemporal behaviors.
• The concept underpins novel states including oscillatory order, glassy dynamics, and defect-mediated limit cycles, opening avenues for advanced research in materials science and network theory.

Nonreciprocal frustration denotes a class of phenomena in which asymmetric interactions, directional propagation, or directed signed relations generate conflicting objectives that cannot be satisfied simultaneously. In the most explicit formulations, it is the dynamical analogue of geometric frustration: static accidental degeneracy of reciprocal systems is replaced by manifolds of marginal orbits, time-dependent ordered states, or persistent cycling in systems without a conventional energy functional. The term is also used in broader settings, including directed signed networks, dissipative quantum spins, and driven metamaterials, where incompatibility is encoded in reciprocity structure, cavity-mediated phase shifts, or space-time modulation rather than in reciprocal bond geometry alone (Hanai, 2022).

1. Conceptual definition and relation to geometric frustration

In reciprocal frustrated magnets, incompatible pairwise constraints prevent all interaction energies from being simultaneously minimized. A standard example is the triangular transverse Ising model with Hamiltonian

$$\mathcal{H}=h(\sigma_x^1+\sigma_x^2+\sigma_x^3)+J(\sigma_z^1\sigma_z^2+\sigma_z^2\sigma_z^3+\sigma_z^1\sigma_z^3),$$

for which the antiferromagnetic case $J>0$ is frustrated because on a triangle it is impossible to satisfy all three pairwise Ising interactions simultaneously, whereas $J<0$ is non-frustrated (Rao et al., 2013).

Hanai formulates a direct analogy between this reciprocal notion and non-reciprocal dynamics. In dissipative XY systems,

$$\dot\theta_i=-\sum_{j=1}^N J_{ij}\sin(\theta_i-\theta_j),$$

the reciprocal case $J_{ij}=J_{ji}$ is gradient descent on

$$E(\theta)=-\sum_{i,j}J_{ij}\cos(\theta_i-\theta_j),$$

while the perfectly non-reciprocal limit $J_{ij}=-J_{ji}$ has no such energy description. The key statement is that geometrically frustrated reciprocal systems and anti-symmetrically coupled non-reciprocal systems both exhibit “accidental degeneracies”: in the former they are degenerate ground-state manifolds, in the latter they are manifolds of marginal orbits with zero Lyapunov directions. The Liouville-type theorem for the anti-symmetric case yields phase-space-volume conservation and

$\sum_{i=1}^N \lambda_i=0,$

which explains why neutral trajectories rather than attractors generically appear (Hanai, 2022).

A sharper distinction is made in dissipative quantum spins. There, geometric frustration is defined as incompatible static energy-minimization constraints, whereas nonreciprocal frustration arises when dynamical objectives are mutually incompatible because the interaction from $i\to j$ is not equal to the interaction from $j\to i$. In that language, one subsystem’s relaxation tendency becomes another’s drive, so there is no common fixed point satisfying all local dynamical objectives (Lyu et al., 8 Aug 2025).

2. Nonreciprocal Ising and XY models as minimal dynamical realizations

A central microscopic realization is the two-lattice nonreciprocal Ising model evolving under Glauber dynamics. Each spin $J>0$0, $J>0$1, has local energy

$J>0$2

The perfectly nonreciprocal case is

$J>0$3

Then one lattice tends to align the paired spins while the other tends to anti-align them. The frustration is therefore not equilibrium frustration on a single Hamiltonian landscape, but a non-Hamiltonian contradiction generated by $J>0$4 and the associated violation of detailed balance (Blom et al., 2024).

The exact thermodynamic-limit evolution equations track both global order,

$J>0$5

and local order,

$J>0$6

The main microscopic result is that coherent oscillations of the global magnetizations are controlled by a critical nearest-neighbor correlation threshold. In the Bethe-Guggenheim approximation, linearization around the disordered state gives a Hopf bifurcation when

$J>0$7

Thus the onset of oscillations is not set directly by $J>0$8 or $J>0$9 alone; it occurs when local order becomes sufficiently strong that the nonreciprocal inter-lattice coupling can no longer be accommodated by a static state. During the oscillatory phase, local order remains high, so the state exhibits nontrivial spatiotemporal correlations rather than simple global disorder (Blom et al., 2024).

The same dynamical logic already appears in the anti-symmetric XY problem. For two spins with $J<0$0, the exact solution

$J<0$1

shows that the relative angle is conserved while the center-of-mass angle drifts. The system therefore realizes a family of marginal drifting orbits parametrized by the initial condition, which is the simplest explicit manifestation of dynamical accidental degeneracy (Hanai, 2022).

3. Fluctuation selection, disorder, and glassy behavior

Noise and quenched disorder do not merely broaden nonreciprocal frustrated dynamics; they can select specific states from an orbit manifold. In Hanai’s treatment, fluctuations around marginal orbits generate an emergent entropic force analogous to equilibrium order-by-disorder. For two communities with equal intra-community couplings $J<0$2 and perfectly non-reciprocal inter-community coupling $J<0$3, the weak-noise expansion yields

$J<0$4

with stable fixed points

$J<0$5

Noise therefore lifts the marginality and selects parity-related chiral states. Away from perfect anti-symmetry, the competition between a reciprocal bias and the entropic torque produces a noise-driven bifurcation at

$J<0$6

beyond which a parity-broken time-dependent phase appears (Hanai, 2022).

The same work reports a non-reciprocity-induced spin-glass-like state in a one-dimensional random XY chain. In the asymmetric case $J<0$7, the spatial correlation decays as

$J<0$8

while the temporal autocorrelation decays algebraically,

$J<0$9

with clear aging. The state has short-ranged spatial order, slow nonstationary dynamics, and no long-range nematic order on accessible timescales (Hanai, 2022).

A different disorder mechanism appears in the two-dimensional Ising ferromagnet with quenched nonreciprocal bonds. For each nearest-neighbor pair,

$$\dot\theta_i=-\sum_{j=1}^N J_{ij}\sin(\theta_i-\theta_j),$$0

Here frustration arises because local update rules cannot be derived from a global energy: a spin may lower its local selfish energy while the global energy stays unchanged or increases. The model remains dynamically active at $$\dot\theta_i=-\sum_{j=1}^N J_{ij}\sin(\theta_i-\theta_j),$$1, shows a continuous transition with

$$\dot\theta_i=-\sum_{j=1}^N J_{ij}\sin(\theta_i-\theta_j),$$2

and satisfies the exact bound

$$\dot\theta_i=-\sum_{j=1}^N J_{ij}\sin(\theta_i-\theta_j),$$3

In the disordered phase, rare-region reversals produce stretched-exponential autocorrelation decay,

$$\dot\theta_i=-\sum_{j=1}^N J_{ij}\sin(\theta_i-\theta_j),$$4

while in the ordered phase coarsening crosses over from $$\dot\theta_i=-\sum_{j=1}^N J_{ij}\sin(\theta_i-\theta_j),$$5 to the activated regime

$$\dot\theta_i=-\sum_{j=1}^N J_{ij}\sin(\theta_i-\theta_j),$$6

This establishes a form of nonreciprocal frustration in which athermal activity prevents conventional zero-temperature freezing (Grodzinski et al., 19 Jun 2026).

4. Finite geometry, parity effects, and exceptional points in the kinetic Ising chain

The exactly solvable one-dimensional kinetic Ising model with non-reciprocity shows that frustration can arise even in the absence of any equilibrium geometric frustration. The local field is

$$\dot\theta_i=-\sum_{j=1}^N J_{ij}\sin(\theta_i-\theta_j),$$7

with reciprocal and non-reciprocal parts

$$\dot\theta_i=-\sum_{j=1}^N J_{ij}\sin(\theta_i-\theta_j),$$8

The one-point dynamics,

$$\dot\theta_i=-\sum_{j=1}^N J_{ij}\sin(\theta_i-\theta_j),$$9

shows immediately that left and right neighbors contribute asymmetrically (Weiderpass et al., 2024).

The exact propagator separates three regimes,

$J_{ij}=J_{ji}$0

and finite systems are separated by $J_{ij}=J_{ji}$1-order exceptional points. In the infinite chain, the propagator peak drifts as

$J_{ij}=J_{ji}$2

so non-reciprocity acts as a drift velocity. In open chains, the wave packet reflects and inverts at the boundary, creating negative correlations and long-lived oscillations. In periodic chains, parity becomes decisive: for $J_{ij}=J_{ji}$3,

$J_{ij}=J_{ji}$4

The paper’s more distinctive result is that sufficiently strong non-reciprocity induces parity dependence even when $J_{ij}=J_{ji}$5, because the asymmetric part can make the effective behavior antiferromagnetic-like. Frustration is therefore boundary-condition sensitive and dynamical: finite rings may be unable to realize the preferred correlation pattern uniformly, despite the absence of frustrated loops in the usual equilibrium sense (Weiderpass et al., 2024).

5. Directed signed networks and dyadic reciprocity

In directed signed networks, frustration is generalized from undirected balance theory to a reciprocity-based notion defined already at the dyadic scale. The classical undirected frustration count,

$J_{ij}=J_{ji}$6

is incomplete for directed graphs because it ignores edge direction. The directed formulation instead distinguishes reciprocal positive dyads, reciprocal negative dyads, reciprocal discordant dyads, and single nonreciprocal dyads. The key motif counts are

$J_{ij}=J_{ji}$7

For datasets whose semantics is closer to mutual judgment or liking/disliking, $J_{ij}=J_{ji}$8 reciprocity is the balanced pattern, whereas nonreciprocal dyads, especially single negative dyads, can be viewed as frustrated when the missing return edge indicates lack of reciprocation (Gallo et al., 2024).

The associated reciprocity indicators are

$J_{ij}=J_{ji}$9

together with analogous nonreciprocal fractions and motif $$E(\theta)=-\sum_{i,j}J_{ij}\cos(\theta_i-\theta_j),$$0-scores. To evaluate empirical over- and under-representation, the paper extends the Exponential Random Graph framework to binary, directed, signed networks, using the Signed Directed Random Graph Model and the Signed Directed Configuration Model with both free-topology and fixed-topology variants (Gallo et al., 2024).

Across the MMOG, Honduras villages, and Spanish schools datasets, the dyadic structure is highly reciprocated and dominated by $$E(\theta)=-\sum_{i,j}J_{ij}\cos(\theta_i-\theta_j),$$1 mutual dyads. Under free-topology benchmarks the pattern resembles a directed extension of weak balance. Under fixed-topology benchmarks, however, $$E(\theta)=-\sum_{i,j}J_{ij}\cos(\theta_i-\theta_j),$$2 reciprocal dyads and single negative dyads are both over-represented, while discordant reciprocal dyads and single positive dyads are under-represented. The resulting contradiction shows that a straightforward directed extension of balance theory is insufficient; higher-order frustration analysis should explicitly constrain signed reciprocity, because bidirectional dyads already supply a strongly structured substrate for triadic motifs (Gallo et al., 2024).

6. Dissipative quantum spins and dynamical frustration in metamaterials

A fully quantum realization is provided by three collective spins coupled to a damped cavity. The Lindblad dynamics is governed by

$$E(\theta)=-\sum_{i,j}J_{ij}\cos(\theta_i-\theta_j),$$3

with

$$E(\theta)=-\sum_{i,j}J_{ij}\cos(\theta_i-\theta_j),$$4

After adiabatic elimination of the cavity in the regime $$E(\theta)=-\sum_{i,j}J_{ij}\cos(\theta_i-\theta_j),$$5, the effective couplings are controlled by

$$E(\theta)=-\sum_{i,j}J_{ij}\cos(\theta_i-\theta_j),$$6

Because

$$E(\theta)=-\sum_{i,j}J_{ij}\cos(\theta_i-\theta_j),$$7

when $$E(\theta)=-\sum_{i,j}J_{ij}\cos(\theta_i-\theta_j),$$8, cavity loss produces genuinely nonreciprocal spin-spin interactions (Lyu et al., 8 Aug 2025).

This model exhibits both nonreciprocal and geometric frustration. In the reciprocal limit, the self-organized phases have degeneracies $$E(\theta)=-\sum_{i,j}J_{ij}\cos(\theta_i-\theta_j),$$9 for the unfrustrated SOP, $J_{ij}=-J_{ji}$0 for the partially frustrated pFSOP, and $J_{ij}=-J_{ji}$1 for the fully frustrated FSOP, with exact triangular antiferromagnetic frustration at $J_{ij}=-J_{ji}$2. The new result is that nonreciprocity stabilizes frustrated phases over a finite region around the $J_{ij}=-J_{ji}$3-symmetric points and makes them robust against disorder. Dynamical phases emerge through exceptional points and include a chiral phase with phase-locked $J_{ij}=-J_{ji}$4 offsets and a swap phase in which the system cycles around the six frustrated metastable states, forming a hexagonal limit cycle in the $J_{ij}=-J_{ji}$5 plane. The swap phase shows multiple odd harmonics and critical slowing down, and the proposed experimental platform is a three-component spinor BEC-cavity system (Lyu et al., 8 Aug 2025).

Driven metamaterials realize an allied form of dynamical frustration. The basic oscillator network obeys

$J_{ij}=-J_{ji}$6

Parametric pumping at $J_{ij}=-J_{ji}$7 produces local $J_{ij}=-J_{ji}$8 phase bistability, while a winding pump phase

$J_{ij}=-J_{ji}$9

makes a globally consistent assignment impossible on a ring. The frustration point becomes a topologically protected phase dislocation that propagates unidirectionally with linear-theory speed

$\sum_{i=1}^N \lambda_i=0,$0

Instead of a degenerate static ground-state manifold, the system forms a self-oscillating defect-carrying limit cycle. Tessellating frustrated loops into two dimensions then yields globally synchronized nonreciprocal phase defects (Mahore et al., 1 Jun 2026).

7. Related usages, diagnostics, and critical relevance

The term is not used identically across all nonreciprocal literatures. In parity-mixed superconductors, the central object is the nonreciprocal superfluid density

$\sum_{i=1}^N \lambda_i=0,$1

which extends the London response to

$\sum_{i=1}^N \lambda_i=0,$2

For the geometry with $\sum_{i=1}^N \lambda_i=0,$3, the penetration depth becomes

$\sum_{i=1}^N \lambda_i=0,$4

The paper explicitly interprets parity mixing as a kind of superconducting frustration between even- and odd-parity channels, with the directional correction to stiffness acting as a bulk asymmetry parameter (Watanabe et al., 2021).

Other works use the language more analogically. The all-passive metasurface paper combines a nonlinear electromagnetic diode and mirror-imaged chiral rotators so that the same observed polarization rotation sense occurs from either side. It states that this is a kind of nonreciprocal frustration because the natural reciprocal behavior of chirality is frustrated by directional selection, although the paper does not develop a formal frustration theory with a Hamiltonian (Mahmoud et al., 2014). The electron-hydrodynamics paper is more explicit that it does not define frustration in the many-body-spin sense; there the operational quantity is directional imbalance, quantified by the reflectance asymmetry

$\sum_{i=1}^N \lambda_i=0,$5

generated by magnetic-field-induced geometric hydrodynamic coefficients and odd-in-$\sum_{i=1}^N \lambda_i=0,$6 conductivity (Sano et al., 2021).

At the level of critical phenomena, nonreciprocity is not automatically relevant. For two identical uncoupled critical fields with antisymmetric perturbation

$\sum_{i=1}^N \lambda_i=0,$7

the first-order correction

$\sum_{i=1}^N \lambda_i=0,$8

implies that constant nonreciprocity is relevant when the susceptibility diverges,

$\sum_{i=1}^N \lambda_i=0,$9

By contrast, if one coupled field is subcritical, or if a finite reciprocal coupling reorganizes the critical modes into one critical and one massive sector, weak nonreciprocity is irrelevant. This places the frustration-like competition of asymmetric coupling on a Harris-style footing: it modifies universality only when it couples into a divergent response sector (Lorenzana et al., 22 Sep 2025).

Taken together, these works show that nonreciprocal frustration is best understood as a family of structurally related but formally distinct mechanisms. In its strictest sense, it is a dynamical frustration generated by asymmetric couplings and manifested through marginal orbits, oscillatory order, persistent cycling, or robust frustrated manifolds. In broader usage, it names situations where reciprocity constraints, parity channels, or propagation directions cannot be made mutually compatible, and where the resulting incompatibility becomes measurable through local-order thresholds, motif over-representation, cavity-field trajectories, defect motion, penetration-depth asymmetry, or directional optical response (Hanai, 2022).