---
title: Weak-Measurement Altered Criticality
url: https://www.emergentmind.com/topics/weak-measurement-altered-criticality
type: topic
---

# Weak-Measurement Altered Criticality

Searching arXiv for recent and foundational papers on weak-measurement altered criticality.
Weak-Measurement Altered Criticality denotes a family of phenomena in which weak, weakly induced, or postselected measurements qualitatively reshape the response or universal structure of a quantum system near a critical, near-singular, or defect-controlled regime. In the literature, the phrase does not refer to a single universal construction. It covers, at minimum, four technically distinct settings: finite-coupling regularization of weak-value amplification in single-system weak measurements; weak-measurement control of monitored many-body entanglement transitions; defect-line deformations of critical ground states and conformal data; and postselected or replicated theories in which measurement randomness or trajectory reweighting drives new fixed points or new universality classes [1709.04869] [1903.05452] [2302.04325] [2301.05238].

## 1. Scope and operative meanings

In weak-value protocols, the relevant structure is the near-singular response associated with the weak value
\[
\langle \widehat{A}\rangle_w = \frac{\langle \psi_f | \widehat{A} |\psi_i \rangle}{\langle \psi_f | \psi_i \rangle},
\]
which becomes large as \(\langle \psi_f|\psi_i\rangle\to 0\). Here “altered criticality” refers to the fact that finite interaction strength regularizes the apparent divergence: the pointer response becomes nonlinear, then biased, and eventually quasi-asymptotic rather than unbounded [1709.04869]. In repeated weak-measurement dynamics with post-selection, the analogous non-analyticity is a stability exchange in a non-unitary Kraus map, controlled by the zero crossing of \(\operatorname{Im}(O_{S,w})\), with relaxation time \(\tau(\phi)\propto |\phi-\phi_c|^{-1}\) and critical exponent \(1\) [2511.03352].

In monitored many-body systems, the term refers to changes in the phase boundary, finite-strength thresholds, or dynamical critical manifolds. A one-dimensional random circuit with Gaussian-probe POVMs exhibits a finite critical measurement strength \((\lambda/\Delta)_{\rm crit}=0.30(1)\) at \(p=1\), below which the system remains ergodic for any measurement probability [1903.05452]. In Fibonacci-monitored Ising/Majorana chains, weak Born-rule measurements shift the critical line to \(\tau_x/\tau_{zz}=1/\sqrt{\phi}\) in the weak-measurement regime, while preserving the long-time weak-self-dual nonunitary universality class with \(c_{\rm ent}\approx 0.793\text{--}0.795\) and \(\nu\approx 1.9\) at Fibonacci times [2605.24086].

In critical ground-state settings, weak or weakly induced measurements are represented as defect insertions in Euclidean spacetime. In the critical Ising chain, ancilla-assisted measurement generates a nonunitary defect line at \(\tau=0\) that couples to \(\sigma\), \(\varepsilon\), or bilocal products thereof, leading either to continuously shifted exponents or to induced order \(\langle Z\rangle_{\tilde s}\sim u^{2/7}\) [2302.04325]. In Tomonaga–Luttinger liquids, arbitrarily weak local measurements over extended regions generate a boundary perturbation \(\cos 2\varphi\), relevant for \(K<1\), while generic outcome ensembles yield replica locking for \(K<1/2\) [2207.09476].

A further usage concerns postselected or replicated criticality. Post-selected measurement-induced transitions reweight rare trajectories and can change universality class outright, with \(\nu=2.1(5)\) and \(c_{\mathrm{eff}}=-0.40(6)\) in a forced-measurement random circuit, matching random tensor networks rather than the standard Born-rule MIPT [2603.15744]. In tricritical and critical Ising ground states under weak measurement, the intrinsic randomness of outcomes drives a finite-coupling measurement-dominated RG fixed point with multifractal exponents, logarithmic factors, independent \(c_n^{(\mathrm{eff})}\), and an effective boundary entropy \(S_{\mathrm{eff}}\) [2409.02107].

## 2. Weak values, near-singular response, and finite-coupling regularization

The canonical single-system framework is the von Neumann indirect measurement scheme, with interaction
\[
\widehat U=e^{-ig\widehat A\otimes \widehat P},
\]
pre-selection \(|\psi_i\rangle\), post-selection \(|\psi_f\rangle\), and a pointer degree of freedom initially prepared in a state \(|\phi(q)\rangle\) [1709.04869]. In the ideal weak limit and for real weak values, the postselected pointer shift is linear,
\[
\langle\Psi_f|\widehat Q|\Psi_f\rangle=|z|^2g\langle \widehat A\rangle_w,
\]
with \(z=\langle\psi_f|\psi_i\rangle\). The central result is that the nominal condition \(g^2\ll 1\) is not sufficient when \(\langle \widehat A\rangle_w\) is anomalously large. The second-order correction vanishes, the next nontrivial term scales as \(g^3\), and finite coupling “determines the range of weak values that one is able to extract” [1709.04869].

Experimentally, for heralded single photons and birefringent-crystal couplings \(g\simeq 0.16\), the first-order weak-value regime survives only over \(\langle\widehat{\Pi}_{H,V}\rangle_w\in[-1.5,2.5]\). For stronger couplings \(g\simeq 0.40\text{--}0.45\), the linear regime contracts to \([-0.7,1.7]\), followed by a third-order biased regime and then a nonlinear regime in which \(\langle \widehat X\rangle\) and \(\langle \widehat Y\rangle\) are “basically constant” and the weak value can no longer be extracted [1709.04869]. This is a direct finite-coupling cutoff on anomalous amplification.

Repeated postselected weak measurements display a related but distinct critical-like phenomenon. For a retained meter acted on by
\[
\hat K=\braket{\psi_f|\psi_S}\left[\hat I-igt\,O_{S,w}\hat O_A\right],
\]
the asymptotic state is selected by the dominant Kraus eigenvalue modulus. The dominant modulus depends on \(\operatorname{Im}(O_{S,w})\), so the stable fixed point switches when \(\operatorname{Im}(O_{S,w})=0\). The resulting discontinuity is not a thermodynamic phase transition and not an exceptional point; it is a dynamical stability exchange in a normalized non-unitary map, with universal exponent \(\nu=1\) under the stated weak-coupling assumptions [2511.03352].

Metrological analyses sharpen the interpretive boundary of these near-singular regimes. Weak-value amplification can generate large conditioned responses near nearly orthogonal post-selection, but the total Fisher information obeys
\[
F_{tot}=p_d Q_d+(1-p_d)Q_r+F_p \le Q_j,
\]
so the amplified signal is typically offset by the small post-selection probability [1310.5302]. This suggests that weak-measurement altered criticality in the weak-value setting is better understood as regulated susceptibility or information redistribution than as a generic gain in ultimate precision.

## 3. Weak measurements in monitored many-body dynamics

In hybrid random circuits, weak measurement introduces a second relevant control parameter beyond the measurement rate. For a Gaussian-pointer POVM with interaction \(H_{\rm int}=\lambda \sigma_z p\), the natural dimensionless strength is \(\lambda/\Delta\), and the system undergoes an entanglement transition between a volume-law ergodic phase and an area-law low-entropy phase only if the measurement strength exceeds a nonzero threshold [1903.05452]. At \(p=1\), finite-size scaling of the half-chain entropy gives
\[
\left(\frac{\lambda}{\Delta}\right)_{\rm crit}=0.30(1),\qquad \gamma=1.38(7),\qquad \nu=1.96(2),
\]
with independent variance-peak extrapolation yielding \((\lambda/\Delta)_{\rm crit}=0.304(3)\) [1903.05452]. Weak measurements therefore alter criticality by making measurement strength itself a genuine critical variable.

Not every weak-measurement protocol changes the universality class. In Haar/dual-Haar hybrid random circuits, replacing strong projective measurements by either an effectively infinite-outcome Gaussian-pointer POVM or a two-outcome softened projector shifts the critical rate \(p_c\) but leaves the universal data consistent, within numerical accuracy, with the projective MIPT. Representative estimates are \(c_{\mathrm{eff}}=0.25(3)\) for the discrete Gaussian pointer model at \(\lambda/\Delta=1\), \(c_{\mathrm{eff}}=0.26(2)\) for the softened-projector model at \(\Lambda=0.45\), \(z\approx 1\), and \(x_1^{\mathrm{typ}}\approx 0.12\text{--}0.14\) across all protocols [2404.02968]. This is a case in which weak measurement alters the phase boundary but apparently not the infrared fixed point.

A more structured alteration occurs in Fibonacci-monitored Ising/Majorana chains. The measurement layers follow the Fibonacci word \(w_\infty=10110101\cdots\), with \(X\)-measurement and \(ZZ\)-measurement layers occurring at asymptotic frequencies controlled by the golden ratio \(\phi\) [2605.24086]. For post-selected trajectories, dynamical self-duality at Fibonacci times places the critical line at
\[
\frac{\tau_x}{\tau_{zz}}=\frac{1}{\phi}.
\]
For Born-rule weak measurements, statistical self-duality instead gives
\[
\frac{\tau_x}{\tau_{zz}}=\frac{1}{\sqrt{\phi}}
\]
in the weak-measurement regime, while the strong-measurement limit approaches the Floquet/projective ratio \(1\) [2605.24086]. The asymptotic universality class along the Born weak-measurement line remains the weak-self-dual nonunitary one, with \(c_{\rm ent}=0.793(1)\) and \(\nu=1.92(5)\), but the critical manifold and intermediate-time entanglement dynamics are quasiperiodically deformed [2605.24086].

## 4. Defect-line deformations of critical ground states

For critical ground states, weak measurement is naturally encoded as a defect localized at a Euclidean time slice. In the critical transverse-field Ising chain, ancilla-assisted measurement with weak entangling parameter \(u\ll 1\) produces conditional states of the form
\[
\ket{\psi_{\tilde s}}=\frac{1}{\sqrt{\mathcal N}\,U' e^{-H_m/2}\ket{\psi_c},
\]
where \(H_m\) contains both linear and bilocal terms built from the system operator used in the ancilla gate [2302.04325]. In cases where \(H_m\) couples to the energy operator \(\varepsilon\), the order-parameter scaling dimension shifts continuously,
\[
\Delta_\sigma(u)=\frac18\left(1+2\kappa u^2 m_{\rm eff}\right),
\]
so the spin-spin correlator acquires an \(O(u^2)\) exponent renormalization [2302.04325]. In cases where \(H_m\) couples linearly to \(\sigma\), the induced line field is relevant and generates
\[
\langle Z\rangle_{\tilde s}\sim u^{2/7}.
\]

The same defect logic appears in critical Tomonaga–Luttinger liquids. For a translation-invariant no-click outcome, the effective action becomes
\[
s_{\rm n.c.}[\varphi]=s[\varphi]-v\int dx\,\cos[2\varphi(x)],
\]
with RG flow \(dv/d\ell=(1-K)v\) [2207.09476]. Thus arbitrarily weak measurements are relevant for \(K<1\), and the asymptotic post-measurement exponents are altered to
\[
\langle \nabla \hat \phi(0)\nabla \hat \phi(x)\rangle_{\rm n.c.}\sim -x^{-2/K},\qquad
\left\langle e^{i[\hat \theta(x)-\hat \theta(0)]}\right\rangle_{\rm n.c.}\sim x^{-1/K}.
\]
For generic outcome ensembles, nonlinear observables are described by a replica action with inter-replica locking term \(\cos 2(\varphi_\alpha-\varphi_\beta)\), relevant for \(K<1/2\), leading to replica-exchange symmetry breaking in the infrared [2207.09476].

A critical qualification is that critical-looking observables need not imply genuine criticality. For massless free fermions under weak measurement of staggered density,
\[
|\psi_M\rangle = \frac{e^{-\beta H_M}|\psi_{\rm GS}\rangle}{\|e^{-\beta H_M}|\psi_{\rm GS}\rangle\|},
\qquad
H_M=\sum_m (-1)^m c_m^\dagger c_m,
\]
the post-measurement state retains \(1/(x-y)\) correlations and logarithmic entanglement entropy, but the entanglement spectrum obeys
\[
\tanh^2 \pi s = \left( \frac{\tanh \pi s_0}{\cosh 2\beta} \right)^2 + \tanh^2 2\beta,
\]
which opens a finite entanglement gap
\[
\Delta=\frac{4}{\pi}\beta
\]
for any nonzero \(\beta\) [2411.13705]. The single-interval entanglement Hamiltonian becomes gapped and long-ranged, not local and gapless as in a conformal ground state. This directly refutes the inference that power-law correlations plus logarithmic entropy are sufficient to establish post-measurement criticality [2411.13705].

## 5. Post-selection, replica structures, and measurement-dominated fixed points

In \((2+1)d\) criticality, postselected weak measurement maps to a boundary or defect perturbation of the \(O(N)\) Wilson–Fisher fixed point. For positive defect mass \(\varepsilon>0\), the monitored state flows to ordinary boundary criticality, with defect order-parameter correlator
\[
\mathrm{tr}\{\rho_P^D\, \hat{\phi}(0)\cdot \hat{\phi}(\mathbf{x})\} \sim \frac{1}{|\mathbf{x}|^{2\Delta_\phi^b}},
\qquad
\Delta_\phi^b = 1+\frac{2}{3N}+O\!\left(\frac{1}{N^2}\right).
\]
For \(\varepsilon<0\), the same framework yields extraordinary-log criticality,
\[
\mathrm{tr}\{\rho_P^D\, \hat{\phi}(0)\cdot \hat{\phi}(\mathbf{x})\} \sim \frac{1}{(\ln |\mathbf{x}|)^q},
\]
rather than a power law [2301.05238]. The same work also identifies an additional transition in nonlinear observables such as \(\mathrm{tr}\{(\rho^D)^2\}\), with doubled-symmetry breaking and an exact lattice realization whose purity maps to the 2d ferromagnetic Ising model, giving \(p_c^{(2)}=0.178\) [2301.05238].

Post-selected monitored circuits can change universality even more drastically because they reweight trajectories. In a forced-measurement random circuit, the post-selected transition has
\[
\nu=2.1(5),\qquad c_{\mathrm{eff}}=-0.40(6),
\]
and matches the random tensor network universality class, in contrast with the standard Born-rule MIPT [2603.15744]. In the translationally invariant post-selected weak-measurement setting, the paper finds that qutrits support the transition whereas qubits do not, indicating strong sensitivity to onsite Hilbert-space dimension [2603.15744].

Measurement randomness itself can generate new fixed points. For tricritical Ising and critical Ising ground states under weak measurement, the replicated defect theory has a measurement-dominated infrared fixed point
\[
\Delta_*=\frac{\epsilon}{4}+\frac{\epsilon^2}{4}+\mathcal O(\epsilon^3)
\]
in the \(R\to1\) measurement replica limit [2409.02107]. At that fixed point, correlation moments acquire multifractal spectra, logarithmic factors appear in selected correlators, the Rényi entropies are controlled by independent \(c_n^{(\mathrm{eff})}\), and the Shannon entropy of the measurement record contains a universal constant identified with an effective Affleck–Ludwig boundary entropy \(S_{\mathrm{eff}}\) [2409.02107]. This is among the clearest formulations of weak-measurement altered criticality as a new universality class rather than a deformation of an old one.

## 6. Experimental realizations, computational approaches, and interpretive limits

A concrete experimental route is provided by Rydberg chains tuned to Ising and tricritical Ising criticality. Projectively measuring periodic subsets of atoms and postselecting outcome patterns produces post-measurement states whose correlators are governed by the same defect-BCFT logic used for weak measurements [2506.21963]. The protocol identifies \(\sigma\)-type and \(\varepsilon\)-type measurement patterns: for Ising, \(\sigma\)-type patterns yield \(\Delta_\sigma=\Delta_\varepsilon=2\), while symmetry-protected \(\varepsilon\)-type patterns keep \(\Delta_\varepsilon=1\) and shift \(\Delta_\sigma\) continuously, with representative values \(\Delta_\sigma\approx 0.2199\), \(0.26968\), and \(0.4101\) for different periodic measurement choices [2506.21963]. The most experimentally favorable post-selection sectors can occur with probabilities of order \(10\%\) in chains with order \(100\) sites [2506.21963].

Weak, ancilla-assisted measurement also alters non-Ising criticality. In a one-dimensional DQCP-analogue chain, \(Z\)-type weak measurement generates an asymmetric restructuring of entanglement across the zFM–VBS transition. For the \((\downarrow\downarrow)\) trajectory, the bipartite entanglement entropy increases strongly for \(K<K_c\) and decreases weakly for \(K>K_c\), while the correlation length develops a growing discontinuity \(\Delta\xi\) at the pseudocritical point as the bond dimension \(\chi\) increases [2603.05436]. The authors argue that this points to a weak first-order phase boundary in the thermodynamic limit, whereas weak \(X\)-type measurement leaves the DQCP-like scaling \(S(\chi)=\frac{c}{6}\ln\xi(\chi)\) with \(c\approx 1\) essentially intact in the weak-measurement regime [2603.05436].

On the computational side, measurement-altered criticality is difficult because postselection and full-state reconstruction scale poorly with system size. A recent proposal therefore recasts the problem as learning the conditional distribution of local reduced density matrices \(p(\rho\mid s_{[i]})\) using a physics-preserving conditional diffusion model [2412.01513]. The application is to ancilla-assisted measurement-altered Ising criticality, where the output is a one- or few-qubit density matrix conditioned on a local measurement window. The proposal is motivated by locality and by the need to access nonlinear observables such as entanglement, but it does not yet provide full quantitative benchmark results [2412.01513].

A final interpretive limit is that not every sharp response or postselected singularity constitutes criticality in the statistical-mechanical sense. In weak-value metrology and in the weak-measurement reinterpretation of Rabi and Ramsey resonances, the enhanced response arises from interference and denominator suppression, not from a diverging correlation length or a new fixed point [1310.5302] [2509.23685]. Conversely, some monitored and postselected systems exhibit altered criticality without changing asymptotic universality, as in Fibonacci-monitored Ising chains or weak-measurement MIPTs in Haar random circuits [2605.24086] [2404.02968]. The literature therefore uses the same phrase for at least three distinct mechanisms: regularization of a near-singular response, deformation of a critical manifold at fixed universality, and emergence of genuinely new fixed points or universality classes.

Source: https://www.emergentmind.com/topics/weak-measurement-altered-criticality