{"id":"40df44d2-6772-4967-80f0-eb25a5c857ff","arxiv_id":"2509.01290","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":3,"one_line_summary":"The paper attempts to build an epsilon-bounded interaction-free Wigner's-friend paradox and a three-box noncontextual inequality, but the explicit circuit fails and the core stability lemma is unproven.","lead":"A quantum-foundations preprint claims a new Wigner's-friend-style paradox based on interaction-free measurements, plus a disturbance-robust three-box inequality. The central construction appears internally inconsistent, and the key stability assumption is not actually derived, so the paper does not establish its claims.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Appendix A's explicit encoding makes both Dark flags imply C=1: Eq. (9) gives bA=CA, so WA=D -> CA=1 -> C=1, matching WB=D -> C=1. No contradiction; flipping bA makes Dark-Dark probability zero.","rationale":"The reader's weakest_assumption is exactly on target: the CLF no-go theorem collapses at the level of the explicit construction because the stated encoding cannot simultaneously produce a nonzero Dark-Dark event and contradictory coin inferences. The inconsistency is internal, not a mere deviation from consensus. The ABL calculation in Appendix E is correct and the three-box algebra is fine, but those support only the secondary three-box paradox, not the primary CLF claim. The hand-waving in Appendix C does not repair the zero-probability or no-contradiction failure. Since the central claim is unsupported as written, the reject verdict stands; no adjustment is needed.","tokens_in":11103,"tokens_out":3864,"duration_ms":47696,"concrete_test":"Recompute the protocol in Appendix A symbolically or with a few lines of QuTiP. Prepare C=|+>, apply U of Eq. (9), then the IFM oracles with Dark iff bomb=1. Compute P(WA=D,WB=D) and the posterior probabilities P(C=0|D,D) and P(C=1|D,D). (i) With bA=CA, bB=X(CB), verify P(D,D)=1/2 and P(C=1|D,D)=1, P(C=0|D,D)=0 — no contradiction. (ii) With bA=NOT CA, compute P(D,D)=0 because CA=0 and CB=|-> are incompatible under U. If both hold, Theorem 1's construction fails exactly as the reader states.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The decisive load-bearing assumption is the claimed inference loop for Theorem 1, resting on the explicit encoding in Appendix A. Section 3 sets bA=CA and bB = X-basis value of CB, with U in Eq. (9): |0>_C -> |0>_CA|+>_CB and |1>_C -> |1>_CA|->_CB. The IFM oracle writes Dark iff the bomb is |1>. Therefore WA=D iff CA=1 iff C=1, and WB=D iff CB is |-> iff C=1. On the initial state |+>_C, both flags Dark occurs with probability 1/2 and entails C=1; there is no contrary inference and no logical collision. Figure 2's chain WA=D -> C=0 contradicts the stated bA=CA: it would require bA to be NOT CA. If one redefines bA=NOT CA to restore the C=0 chain, then WA=D requires CA=0, while WB=D requires CB=|->; no input C in the range of Eq. (9) satisfies both, so P(WA=D,WB=D)=0. The 'straightforward calculation' of Appendix A is therefore either wrong or cannot yield both nonzero postselection and contradiction. Continuity arguments in Appendix B cannot rehabilitate a probability-zero event, and Appendix C concerns a different three-box inequality, not the CLF construction. Thus Theorem 1, the paper's primary claim, is unsupported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims two main results: (i) a new 'Counterfactual Local Friendliness' (CLF) no-go theorem, in which two Wigners postselect on two interaction-free Dark flags and derive contradictory certainties about a single coin under assumptions (Q), (S), (C), and (IF-epsilon); and (ii) an epsilon-stable three-box noncontextual inequality P(A)+P(B) <= 1 + K*epsilon that quantum theory violates for arbitrarily small epsilon. Appendices sketch extensions to GHZ, Peres-Mermin, Leggett-Garg, and local-friendliness inequalities. The ABL computations for the three-box probabilities are correct, and inequality (7) is algebraically valid conditional on its epsilon-stability premise. However, the explicit unitary encoding in Appendix A does not produce the claimed CLF inference loop: with bA=CA and bB the X-basis value of CB, both Dark flags imply C=1, so there is no logical collision; modifying the encoding to force a collision makes the postselected event have zero probability. The three-box inequality's epsilon-stability premise is also not justified for the A/B context switch. Theorem 1 and the quantitative paradox are therefore unsupported.","tokens_in":11434,"tokens_out":7584,"duration_ms":88079,"significance":"If the CLF paradox were correct, it would be a notable conceptual advance, isolating counterfactuality rather than measurement disturbance as the source of Wigner's-friend contradictions, and the epsilon-bounded formalization would be useful. The paper is transparent in stating assumptions and includes some correct technical pieces: the ABL derivation in Appendix E is straightforward and correct, and the bound (7) does follow from its premises. These local strengths do not rescue the central claim. Because the main theorem rests on an encoding that either yields no contradiction or yields zero postselection probability, and because the three-box inequality's key stability premise lacks a valid derivation for the specified experimental switch, the manuscript is not in a publishable state. The novelty claim for CLF cannot be accepted.","major_comments":[{"comment":"The explicit encoding does not realize the claimed inference loop. With U in Eq. (9), CA equals C in the computational basis and CB is |+>/|-> according to C. Setting bA=CA makes WA=D imply C=1; setting bB to be the X-basis value of CB (with bB=1 for |->) makes WB=D imply C=1 as well. Both flags Dark occurs with probability 1/2 and entails C=1, so there is no contradiction. Figure 2's branch 'WA=D => C=0' would require bA = NOT CA, which contradicts the stated bA=CA. This is not a presentation issue: Theorem 1 rests entirely on this inference loop.","section":"Section 3 / Appendix A / Fig. 2"},{"comment":"If one redefines bA=NOT CA to try to restore Fig. 2, the postselected event becomes empty. The two Dark conditions are CA=0 and CB=|->. But the range of U in Eq. (9) contains only components |0>_CA|+>_CB and |1>_CA|->_CB; no component satisfies both. Hence P(WA=D,WB=D)=0 in the ideal unitary model. Appendix B's O(sqrt(epsilon)) continuity bound cannot make a probability-zero event nonzero at epsilon=0; it only bounds changes around a nonzero value. Thus Theorem 1's 'nonzero probability and contradiction' cannot be achieved by this construction.","section":"Appendix A / Appendix B"},{"comment":"The three-box inequality is conditional on epsilon-stability (Def. 3), and the text claims this follows from epsilon-counterfactuality via data-processing, citing Appendix F. But Appendix C's Lemma 1 only covers contexts that differ by insertion/removal of an IFM module on the same target. In the three-box experiment, contexts A and B differ by which arm (A or B) is probed; the bomb is placed in different physical locations, and the postselected ontic distributions mu(·|A), mu(·|B) can differ for reasons unrelated to the epsilon-disturbance of a single module. No argument is given that TV(mu(·|A),mu(·|B)) <= K' epsilon for this operational switch. Since inequality (7) and its claimed violation (8) depend on this premise, the quantitative paradox is not established.","section":"Section 4.3 / Appendix C"}],"minor_comments":[{"comment":"Definition numbering is inconsistent: Section 1.1 refers to '(IF-epsilon) epsilon-counterfactuality (Def. 2.2)', but the definition in Section 2.2 is labeled 'Definition 2'. Please unify.","section":"Section 1.1 / Section 2.2"},{"comment":"Appendix ordering and cross-references are confusing: E.1 and E.2 appear before the heading 'E Full derivation...', and Section 4.3 refers to 'Appendix F' for the epsilon-stability justification, but the actual justification appears in Appendix C. Correct the references.","section":"Appendix E / Appendix F"},{"comment":"The arrow labels 'WA=D => bA=1 => C=0' and 'WB=D => bB=1 => C=1' conflict with the equations in the text. If the intended encoding differs from the one stated, it should be written out explicitly.","section":"Figure 2"},{"comment":"The 'Caveat' paragraph about earlier drafts speaking of 'two clicks at dark ports' is unexplained and does not address the consistency issue; it should be removed or clarified.","section":"Section 3, Caveat"}],"recommendation":"reject","confidential_remarks":"The central construction is internally inconsistent: the paper's own Appendix A contradicts Figure 2, and the alternative encoding that would restore the contradiction gives zero postselection probability. I do not see a local fix that preserves the stated protocol; a fundamentally different construction would be needed. The three-box bound is a conditional algebraic result, but its physical premise is not established for the described experiment."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Thanks for the report. I read the paper against the stress-test note, and the note is right. The primary result, Theorem 1, fails at the level of its own explicit encoding. Appendix A's Eq. (9) sets bA = CA and bB as the X-basis of CB. With U: |0>C -> |0>_CA |+>_CB and |1>C -> |1>_CA |->_CB, both Dark flags imply C=1. So the postselected event, which does occur with probability 1/2 on |+>_C, entails no contradiction. If you redefine bA to force C=0, the Dark-Dark joint probability drops to zero. The paper's Figure 2 chain WA=D -> C=0 contradicts Eq. (9). Appendix A says a 'straightforward calculation' shows p>0 and the inference loop, but no calculation is given, and the continuity argument in Appendix B can't make a zero-probability event nonzero. That is a load-bearing flaw.\n\nWhat the paper does well: the ABL calculations in Appendix E are correct, and the three-box inequality PA+PB ≤ 1+Kϵ is algebraically valid given the epsilon-stability premise. The authors are also honest in Appendix K that this is a specialization of known noncontextuality inequalities with an epsilon slack. The formalization of epsilon-counterfactual IFM as a trace-distance bound on the bomb is a useful operational notion, and the Zeno scaling discussion in Appendix G is sensible.\n\nSoft spots beyond the CLF failure: the epsilon-stability premise (Def. 3) is justified in Appendix C only as a 'proof idea' sketch that assumes the connection it needs to prove — that bomb disturbance bounds the change in postselected ontic distributions. The paper itself flags this as a sketch, so it can't carry a no-go theorem. The later appendices (GHZ, Peres–Mermin, Leggett–Garg, LF inequality) are templates, not proofs; they add length without adding support.\n\nWho is this for? A reader interested in Wigner's-friend or interaction-free measurement might find the epsilon-counterfactual framing worth a skim, but the core paradox is not established, and the inequality is a minor extension. I would not want to spend referee time on this in its current form. If the authors can provide an actual unitary realization that produces the claimed logical collision with nonzero probability, that would change things — but as it stands, the paper's central claim is unsupported. I'd desk-reject with an invitation to resubmit a corrected version.","headline":"The CLF paradox contradicts its own explicit encoding and the three-box inequality is a known result with an epsilon slack; the paper's main claims are unsupported.","tokens_in":11912,"tokens_out":2996,"would_cite":false,"duration_ms":33187,"reading_group":"maybe","serious_thinker":"no","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P05","81P15"],"pacs":["03.65.Ta"],"model":"deepseek-v4-flash","headline":"A new paradox, Counterfactual Local Friendliness, shows that interaction-free flags with arbitrarily small but nonzero disturbance can force incompatible certainties about a single upstream variable; the same logic yields a three-box inequa","keywords":["counterfactual local friendliness","interaction-free measurement","Wigner's friend","three-box paradox","epsilon-counterfactuality","contextuality","disturbance bound","single-world facts"],"falsifier":"Evaluate the explicit Appendix A circuit with the Section 3 assignments bA=CA and bB=X-basis of CB: compute the probability of {WA=D, WB=D} and the coin value each Dark flag entails. The claim collapses if that probability is zero, or if both flags entail the same C value under the printed Eq. (9).","tokens_in":10960,"feed_emoji":"⚛️","tokens_out":13437,"duration_ms":156210,"temperature":0.7,"pith_summary":"The paper argues that the defining puzzle of interaction-free measurement is not the residual disturbance it leaves on the probed object, but the way agent-level facts collide in a single-world narrative. It builds a Wigner's-friend scenario—an outside observer models a friend's measurement unitarily—with two friends who each use an ε-counterfactual IFM oracle, a unitary flag that certifies a bomb state while changing it by at most ε in trace distance, and the outside Wigners post-select on both flags being Dark. Under universal unitarity, single-outcome facts, cross-agent consistency, and ε-counterfactuality, the two Dark flags entail opposite values for one upstream coin, a logical collision named Counterfactual Local Friendliness. The same ε framework yields a three-box noncontextual inequality: single-world noncontextual models with exclusivity and ε-stability satisfy PA+PB ≤ 1+Kε, while quantum predictions give PA=PB=1 and violate it for arbitrarily small ε. If the construction is sound, it closes the \"maybe it interacted a little\" escape route and gives a disturbance-robust, quantitative witness of nonclassicality.","feed_headline":"Even epsilon-tiny probes can force a Wigner's-friend logical collision","feed_subtitle":"Interaction-free flags with bounded disturbance close the 'it interacted a little' loophole for Wigner's-friend paradoxes.","key_machinery":"The load-bearing object is the ε-counterfactual IFM oracle: a unitary gadget, described as a mediator qubit passing through H–controlled-Z–H with a forwarded flag, whose Dark outcome certifies that the bomb was live while changing the bomb's reduced state by at most ε in trace distance. Each lab uses such an oracle, and the flags are chained through a coherent isometry U that maps one upstream coin qubit into two lab registers, so the two Dark flags inherit opposite conclusions about the same coin. For the three-box result, the same unitaries are paired with ε-stability, a total-variation bound on how much the post-selected ontic distribution changes when the probe context is switched; that","core_discovery":"The central claim is Theorem 1, the CLF No-Go: under assumptions (Q) universal unitarity, (S) single-outcome facts, (C) cross-agent consistency, and (IF-ε) ε-counterfactuality of the friends' internal modules, there is a unitary-only protocol and a threshold ε0>0 such that for every 0≤ε<ε0 the post-selected event {WA=D, WB=D} has nonzero probability and forces logically incompatible certainties about a single upstream coin C. One chain of reasoning makes C=0 certain; the other makes C=1 certain. The proof uses a coherent coin-to-register isometry and unitary IFM oracles so that all decisive inferences come from interaction-free flags rather than absorptive or projective in-lab measurements.","pith_inferences":["Going beyond the paper: the contradiction lives or dies with the choice of encoder from the upstream coin to the lab registers; a direct calculation of the printed Appendix A mapping would settle whether both Dark flags can occur together and whether they entail opposite coin values.","Going beyond the paper: the visibility ratio Vdec/V0 suggested in the appendix gives a direct experimental route to bounding ε; combining it with the three-box bound turns the inequality into a quantitative test one could run in a single interference experiment.","Going beyond the paper: if the collision is robust for all small ε, the same flag-based encoding should transfer to other pre/post-selected paradoxes and to fault-tolerant settings, where an additive ε budget across rounds could certify nonclassicality without invoking zero back-action."],"forward_implications":["The \"it interacted a little\" escape route is closed: the CLF collision persists for every ε below a threshold, so the paradox cannot be resolved by noting that the flag slightly disturbs the bomb.","Any single-world theory that reproduces the Dark-flag certainties with ε-bounded disturbance must abandon universal unitarity or cross-agent consistency; the three-box inequality puts a quantitative price on the trade-off.","The three-box effect becomes a device-agnostic, disturbance-robust witness: measured PA+PB should approach 2 against a classical cap of 1+Kε even as ε→0.","The contradiction does not depend on exact numerical probabilities: the modal version works with possibility and necessity only, so it remains stable under small errors of order √ε.","The ε-composition and visibility-to-ε relations give a practical route to ε-certified low-dose imaging and to scheduling low-back-action quantum error-correction checks."],"supporting_citations":[{"why":"supplies the original interaction-free measurement scenario that the unitary IFM oracle generalises.","marker":"[1]"},{"why":"provides the Zeno-boosted, high-efficiency IFM implementations that make ε small in practice and motivate the ε parameter.","marker":"[2, 3]"},{"why":"gives the local-friendliness inequality template and its violation that the CLF construction adapts to interaction-free flags.","marker":"[7]"},{"why":"is the prior extended Wigner's-friend analysis with interaction-free record detection that the paper claims to strengthen.","marker":"[11]"},{"why":"supplies the gentle-measurement lemma used to derive ε-stability and the K1ε+K2√δ slack in the inequalities.","marker":"[15]"},{"why":"gives the Fuchs–van de Graaf inequality used to propagate ε-bounded disturbance into O(√ε) changes in certainty.","marker":"[17]"}],"fun_headline_variants":["Interaction-free probes force a Wigner's-friend paradox","Bounded-disturbance probes create Wigner's-friend collision","Epsilon-counterfactual logic clashes in Wigner's-friend setup","Three-box inequality: quantum beats noncontextual even with epsilon","Counterfactual friend paradox survives epsilon-tiny disturbance"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The construction works only if the explicit mapping from the coin to the lab registers (Eq. 9) really yields both a nonzero joint Dark-Dark probability and opposite inferred coin values; with the mapping as printed, both Dark flags can be read as pointing to the same coin value, and changing the mapping to force the contradiction can make the Dark-Dark event impossible.","fun_headline_variants_meta":{"raw":{"variants":["Interaction-free probes force a Wigner's-friend paradox","Bounded-disturbance probes create Wigner's-friend collision","Epsilon-counterfactual logic clashes in Wigner's-friend setup","Three-box inequality: quantum beats noncontextual even with epsilon","Counterfactual friend paradox survives epsilon-tiny disturbance"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000201,"raw_usage":{"total_tokens":1235,"prompt_tokens":783,"completion_tokens":452,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":527,"completion_tokens_details":{"reasoning_tokens":366}},"tokens_in":527,"tokens_out":452,"duration_ms":5625,"temperature":1.0,"reasoning_tokens":366,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T12:41:23.917990+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the explicit Appendix A circuit with the Section 3 assignments bA=CA and bB=X-basis of CB: compute the probability of {WA=D, WB=D} and the coin value each Dark flag entails. The claim collapses if that probability is zero, or if both flags entail the same C value under the printed Eq. (9).","supporting_citations":[{"cited_title":"Quantum mechanical interaction-free measurements,","cited_arxiv_id":null,"evidence_quote":"supplies the original interaction-free measurement scenario that the unitary IFM oracle generalises."},{"cited_title":"Relational Analysis of the Frauchiger–Renner Paradox and Interaction-Free Detection of Records from the Past,","cited_arxiv_id":null,"evidence_quote":"is the prior extended Wigner's-friend analysis with interaction-free record detection that the paper claims to strengthen."},{"cited_title":"Cryp- tographic distinguishability measures for quantum-mechanical states,","cited_arxiv_id":null,"evidence_quote":"gives the Fuchs–van de Graaf inequality used to propagate ε-bounded disturbance into O(√ε) changes in certainty."}],"review_version":1}