Causal Stinespring theorem for local quantum field measurements: the operational characterization of physically implementable local operations in QFT

In non-relativistic quantum theory the Stinespring dilation theorem guarantees that every completely positive map on a system arises from a unitary interaction with a probe. In quantum field theory (QFT), naively local operations can signal between spacelike separated regions (Sorkin's 'impossible measurements'), so a causal version of the dilation theorem is required. The source paper (Simmons, Papageorgiou, Christodoulou, Brukner, arXiv:2511.21644) proved that local S-matrices obeying factorisation conditions — continuous additivity or the weaker Hammerstein property — exclude superluminal signalling and retrocausality, derived causality conditions for the induced Kraus operators in an exactly solvable pointer model, introduced effective Kraus operators localised to the causal past of the measured region, and showed that measurement accuracy is bounded below by the field's retarded propagator. It explicitly leaves open the central question: a constructive, necessary-and-sufficient characterization (a 'causal Stinespring theorem') of the unital completely positive maps on local QFT algebras that admit dilations to local, factorisation-obeying probe–field unitaries, and it records the limitations of its own special-case analysis that stand between the present results and such a theorem — generality beyond the quantum-control ansatz and the exactly solvable linear pointer coupling, rigorous existence and uniqueness of the effective Kraus operators, the expected forced destructiveness of measurements of non-local field functionals, and the simplified model setting (trivial pointer dynamics, single scalar field), with independently varying masses and a dynamical probe left to future work. Later 2026 work sharpens the landscape (causal channels are nowhere dense in local channels; within the Fewster–Verch framework the implementable set provably differs from the causal set and its approximate characterization is undecidable for Weyl channels) but does not construct the theorem. The problem remains open.

Background

Quantum measurement theory is well understood for non-relativistic systems. A physically admissible change of state is described by a completely positive (CP) map, and the Stinespring dilation theorem guarantees that every CP map on a system can be realized by a unitary interaction with an auxiliary probe system followed by discarding the probe; the induced state changes are expressible through Kraus operators (operators $K_i$ whose action on a density matrix $\rho$ is $\rho \mapsto \sum_i K_i \rho K_i^\dagger$), and in the Heisenberg picture the relevant channels are unital CP (UCP) maps, which preserve the identity operator.

Quantum field theory complicates this picture radically. An operation that looks perfectly local — for example, projecting the field in a bounded spacetime region onto a state, or applying a local unitary kick — can be exploited by spacelike separated agents to signal faster than light. Such operations are called 'impossible measurements', after Sorkin, who first exhibited them. Algebraic QFT assigns to each spacetime region $R$ a von Neumann algebra $\mathcal{A}(R)$ of local observables; spacelike separated algebras commute (microcausality), but — unlike finite-dimensional systems — they are not tensor factors of one another, which is why the finite-dimensional characterization of causal operations does not transfer. In finite dimensions, Beckman–Gottesman–Nielsen–Preskill and Eggeling–Schlingemann–Werner proved that semicausal operations (those that cannot signal between tensor factors) coincide exactly with semilocalizable ones (implementable by local unitaries on the factors, possibly with shared ancillas); this is the finite-dimensional prototype of a causal dilation theorem, and no analogue for local QFT algebras is known.

Several programs have approached a measurement theory for QFT. The Fewster–Verch (FV) framework realizes measurements by coupling the target field to a probe quantum system through an interaction localized in a spacetime region and then measuring the probe; it guarantees locality and covariance, and it excludes Sorkin-type operations at the level of the apparatus ('impossible measurements require impossible apparatus', in the result of Bostelmann–Fewster–Ruep). Within and around this framework it is known that every local observable of a linear scalar field is asymptotically measurable; that causal state updates exist for smeared fields $\phi(f)$ (the field operator integrated against a test function $f$); and that Gaussian measurements of products of smeared fields are necessarily signalling. Oeckl constructed causal measurements from the spectral decomposition of field operators, proved their causal transparency (measurements in one region do not disturb expectation values in spacelike regions) and spacetime compositionality, and, with Zampeli, pursued a compositional route via the positive formalism (a real-vector-space generalization of quantum theory in which the composition of processes in arbitrary spacetime regions is built into the algebraic framework). Jubb, and separately Fewster–Jubb–Ruep, explicitly called for the missing structural result: a dictionary between update maps and concrete probe models 'analogous to Stinespring's Dilation Theorem, but with the added restrictions of locality and causality on the unitary map' — in short, a causal Stinespring theorem.

The source paper of the present problem (Simmons, Papageorgiou, Christodoulou, Brukner) works in the local S-matrix formalism of causal perturbation theory: the interaction of a probe with the field is encoded in a unitary scattering operator $S[f]$ depending functorially on a smearing function $f$ that specifies where and how the coupling is switched on. The formalism carries a hierarchy of factorisation conditions: causal factorisation $S_{a+b}=S_b S_a$ for disjoint, causally orderable coupling regions; continuous additivity $S[f]=S[f_+]S[f_-]$ whenever the support of $f$ is split by an arbitrary Cauchy surface into future and past parts; and the weaker Hammerstein property $S[f_1+f_2+f_3]=S[f_1+f_2]\,S[f_2]^{-1}\,S[f_2+f_3]$, with $f_1$ and $f_3$ separated by a hypersurface and $f_2$ arbitrary, used in causal perturbation theory for very singular interactions. The paper proves these conditions sufficient to exclude Sorkin's impossible operations, including retrocausal ones. It realizes $S[f]$ by an exactly solvable model — a von Neumann-type measurement of $\phi(f)$ through linear coupling to a pointer with trivial internal dynamics ($H_{\text{pointer}}=0$) — and derives the induced field Kraus operators $\Pi_q^\psi[f]$, a compositional convolution identity for them under a quantum-control ansatz (the S-matrix first reads a probe observable and then applies a field unitary conditioned on its value, a continuum analogue of a controlled gate), the notion of effective Kraus operators localised (in a suitable limiting sense) to the causal past of the region being updated, and a fundamental accuracy bound: the sharpness with which $\phi(f)$ can be measured is bounded below by the square root of the retarded propagator $\Delta_r(f,f)$, i.e. by the causal connectivity of the coupling region. It also concludes that genuinely non-local field functionals (e.g. $\phi(f)^2$, which cannot be written as a sum of spacelike separable parts) admit only destructive causal measurements, in line with the finite-dimensional constraints on measuring non-local variables found by Popescu and Vaidman.

What the paper does not do — and states it cannot yet do — is characterize which maps are implementable; it proves special cases and 'hints' only. Two 2026 developments sharpen the shape of the target without closing it. Simmons (2026) proved that the causal normal local channels (normal meaning continuous in the ultraweak operator topology) form a nowhere dense subset of all normal local channels, so the implementable set is topologically small and cannot be found or characterized by perturbation arguments. Mandrysch–Simmons–Navascués (2026), working inside the FV framework for free scalar fields, proved that all diagonal Weyl instruments (random-displacement channels of the field operators) are FV-realizable — which allows the heralded, probabilistic realization of arbitrary Weyl instruments and deterministic non-demolition measurements — but also constructed causal channels admitting no approximate FV realization (so the causal and FV-realizable sets are genuinely different), and showed that deciding (approximate) implementability of Weyl channels within networks of FV schemes is undecidable, by reduction to the undecidability of characterizing quantum correlations. The dilation characterization for the local-S-matrix/factorisation route of the source paper therefore remains entirely open.

Problem Statement

State and prove a causal Stinespring theorem for local operations in quantum field theory: characterize exactly which unital completely positive maps — or, more generally, quantum instruments, i.e. families of outcome-labelled completely positive maps describing a selective measurement together with its outcome statistics (an announced generalization: the source paper's proven results concern non-selective UCP channels and explicitly exclude selective operations from its locality definition) — on the local algebra of a spacetime region of a QFT admit a dilation to an interaction with a probe, i.e. to a local S-matrix $S[f]$ generated by a unitary coupling between smeared field operators localized in the region and probe degrees of freedom, obeying the factorisation conditions (continuous additivity with respect to arbitrary Cauchy-surface splittings of the coupling support, or the weaker Hammerstein property, together with causal factorisation between disjoint causally orderable regions), such that the induced Kraus operators on the field satisfy the corresponding causality (no-signalling) conditions, in the model class of the source paper (a single real scalar field, a pointer with trivial internal dynamics, the quantum-control ansatz, and a linear pointer coupling). The primary target is the non-selective unital completely positive maps; the analogous characterization of the more general selective quantum instruments is a further strengthening, not required for a resolution. The characterization is required to be constructive — an explicit dictionary from admissible maps to probe models and couplings. The following open directions correspond to limitations the source paper acknowledges in its own special-case results and remain open within the same objective:

1. Generality of couplings. Identify the classes of probe dynamics and interaction Hamiltonians, beyond the quantum-control ansatz (the S-matrix reads a probe observable and then applies a field unitary conditioned on its value) and beyond the exactly solvable linear pointer coupling, for which the compositional Kraus identities (the convolution-type factorisation identity) and the effective-Kraus localisation property hold; the source paper shows the conclusions hold for factorising S-matrices in general, but the Kraus-level identities were established only under these restrictive assumptions.

2. Well-definedness of the effective Kraus operators. Establish the functional-analytic conditions — choices of topologies on the smearing functions, domains, and continuity hypotheses on the map $f\mapsto\Pi[f]$ — under which the effective Kraus operators, i.e. the would-be operators depending only on the causal past of the region they update and defined through the limiting procedure of Cauchy-surface splittings approaching the past lightcone (a limit that cannot be replaced by directly 'chopping' the interaction, which introduces divergences), exist and are unique for general local S-matrices.

3. Non-local observables. Classify the admissible measurement channels for genuinely non-local field functionals — those not reducible to a single smeared field, such as $\phi(f)^2$ — for which the source paper argues that every causal selective update should be destructive.

4. Model generality. Extend the source paper's special-case results — not the mandatory characterization — beyond the simplified setting it analyses (pointer with trivial internal dynamics, $H_{\text{pointer}}=0$; single real scalar field). The source's own stated extension program covers letting the field masses vary independently and reintroducing a nontrivial probe Hamiltonian, i.e. fully dynamical relativistic probes; further directions — multiple and interacting fields, and perturbative as well as non-perturbative interaction regimes — are recorded here as the open-ended frontier of the same objective rather than as separately bounded deliverables attributed to the source.

These components jointly describe one scientific objective — the operational implementability criterion for measurements and operations on quantum fields, with the causal Stinespring theorem as the central deliverable and items 1–4 as the open directions in which the source paper's special-case results remain incomplete. Establishing the central characterization (with its constructive dictionary) resolves the task; items 1–4 are additional open directions that strengthen, but are not required for, a resolution. Since the goal endorsed by the source is, in Jubb's formulation, to construct an explicit dictionary between update maps and specific probe models and 'determine if this is possible in general', an answer may alternatively establish that no such characterization exists — for example a rigorous impossibility or undecidability result for the dilation problem within the local S-matrix/factorisation framework itself, going beyond the known undecidability of approximate implementability for Weyl channels under networks of Fewster–Verch schemes.

References

1. Robin Simmons, Maria Papageorgiou, Marios Christodoulou, Časlav Brukner, Factorisation conditions and causality for local measurements in QFT (2026), arXiv:2511.21644, https://arxiv.org/abs/2511.21644
2. Rafael D. Sorkin, Impossible Measurements on Quantum Fields (1993), arXiv:gr-qc/9302018, https://arxiv.org/abs/gr-qc/9302018
3. David Beckman, Daniel Gottesman, M. A. Nielsen, John Preskill, Causal and localizable quantum operations, Phys. Rev. A 64, 052309 (2001), doi:10.1103/PhysRevA.64.052309, arXiv:quant-ph/0102043, https://arxiv.org/abs/quant-ph/0102043
4. T. Eggeling, D. Schlingemann, R. F. Werner, Semicausal operations are semilocalizable, Europhys. Lett. 57, 782 (2002), doi:10.1209/epl/i2002-00579-4, https://doi.org/10.1209/epl/i2002-00579-4
5. Christopher J. Fewster, Rainer Verch, Quantum fields and local measurements, Commun. Math. Phys. 378, 851 (2020), doi:10.1007/s00220-020-03800-6, https://doi.org/10.1007/s00220-020-03800-6
6. Henning Bostelmann, Christopher J. Fewster, Maximilian H. Ruep, Impossible measurements require impossible apparatus, Phys. Rev. D 103, 025017 (2021), doi:10.1103/PhysRevD.103.025017, arXiv:2003.04660, https://arxiv.org/abs/2003.04660
7. Ian Jubb, Causal state updates in real scalar quantum field theory, Phys. Rev. D 105, 025003 (2022), doi:10.1103/PhysRevD.105.025003, arXiv:2106.09027, https://arxiv.org/abs/2106.09027
8. Christopher J. Fewster, Ian Jubb, Maximilian H. Ruep, Asymptotic measurement schemes for every observable of a quantum field theory, Ann. Henri Poincaré 24, 1137 (2023), doi:10.1007/s00023-022-01239-0, arXiv:2203.09529, https://arxiv.org/abs/2203.09529
9. Emma Albertini, Ian Jubb, Are ideal measurements of real scalar fields causal? (2023), arXiv:2306.12980, https://arxiv.org/abs/2306.12980
10. Robert Oeckl, Spectral decomposition of field operators and causal measurement in quantum field theory, J. Math. Phys. 66, 042302 (2025), doi:10.1063/5.0245368, arXiv:2409.08748, https://arxiv.org/abs/2409.08748
11. Robert Oeckl, Causal measurement in quantum field theory: spacetime, Phys. Rev. D 113, 085002 (2026), doi:10.1103/5gl8-3j6p, arXiv:2511.06566, https://arxiv.org/abs/2511.06566
12. Robert Oeckl, Adamantia Zampeli, Towards local and compositional measurements in quantum field theory (2025), arXiv:2505.10968, https://arxiv.org/abs/2505.10968
13. Jan Mandrysch, Miguel Navascués, Quantum field measurements in the Fewster–Verch framework, Lett. Math. Phys. 115, 115 (2025), doi:10.1007/s11005-025-02001-3, arXiv:2411.13605, https://arxiv.org/abs/2411.13605
14. Robin Simmons, Causality is rare: some topological properties of causal quantum channels (2026), arXiv:2603.25315, https://arxiv.org/abs/2603.25315
15. Jan Mandrysch, Robin Simmons, Miguel Navascués, Causality and realizability of local operations in quantum field theory (2026), arXiv:2607.12976, https://arxiv.org/abs/2607.12976
16. F. Daem, L. Ballesteros Ferraz, A. Zampeli, A. Matzkin, Sensing relativistic quantum fields with minimally perturbing local measurements (2026), arXiv:2607.17920, https://arxiv.org/abs/2607.17920
17. Christopher J. Fewster, Measurement and preparation protocols for quantum field theory on curved spacetimes (2025), arXiv:2508.21426, https://arxiv.org/abs/2508.21426
18. Nadia Koliopoulou, Charis Anastopoulos, Ntina Savvidou, Quantum field theory measurements for relativistic particles (2026), arXiv:2602.14175, https://arxiv.org/abs/2602.14175
19. Sandu Popescu, Lev Vaidman, Causality constraints on nonlocal quantum measurements, Phys. Rev. A 49, 4331 (1994), doi:10.1103/PhysRevA.49.4331, https://doi.org/10.1103/PhysRevA.49.4331
20. Lev Vaidman, Instantaneous measurement of nonlocal variables, Phys. Rev. Lett. 90, 010402 (2003), doi:10.1103/PhysRevLett.90.010402, https://doi.org/10.1103/PhysRevLett.90.010402
21. L. Borsten, I. Jubb, G. Kells, Impossible measurements revisited, Phys. Rev. D 104, 025012 (2021), doi:10.1103/PhysRevD.104.025012, https://doi.org/10.1103/PhysRevD.104.025012