{"id":"8d3bd2d6-2e72-4eef-9351-df7210d63a2d","arxiv_id":"2608.09823","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The authors define Coecke flow lines as branch-independent paths through quantum protocol diagrams that survive every semantics-preserving rewrite down to a bare wire.","lead":"String-diagram rewriting shows when two pictures of a quantum protocol mean the same thing; this paper asks how to follow a single information line through every intermediate picture down to a bare wire. The result gives the old notion of quantum information flow a form that can in principle be checked rewrite by rewrite, and that the authors connect to holographic bit threads in a companion paper.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central formalization rests on the undefined primitive 'apparent through-path': without a graph-theoretic definition of L(D), the boundary-fibre machinery and the local-to-global gluing principle are not well-defined.","rationale":"The reader's weakest_assumption already flagged that 'crosses no visible tensor-product gap' is never made precise at the level of graph or category formalism. My stress-test deepens that concern: the imprecision affects not just the intuitive reading but the actual validity of the local-to-global gluing principle (34), because a global path can intersect a local replacement region in more than one segment unless the notion of apparent through-path is explicitly restricted. This is the load-bearing spot for the paper's central claim of providing an explicit, testable formalization of Coecke flow lines. The defect is real but addressable: one can formalize L(D) in the incidence-graph or hypergraph language and then re-derive the boundary-fibre tables and gluing statements. Because the reader already issued a conditional accept and this concern is a sharpening of the condition rather than a demonstration that the concept is incoherent, I do not move the verdict. I would, however, make the formalization of apparent through-paths and the single-intersection assumption explicit conditions for acceptance, and I would ask the authors to recompute the ZX examples under the chosen formal definition.","tokens_in":39082,"tokens_out":8375,"duration_ms":90070,"concrete_test":"Formalize L(D) on the underlying undirected incidence graph of a ZX diagram, with each spider as a vertex connected to its incident wires, boxes as subgraphs, and tensor factors distinguished by connected components. Then (1) recompute the boundary-fibre tables for spider fusion and the Hopf rule from this definition and compare with Eqs. (52) and (53). (2) Test the four-legged spider counterexample: take P as one Z-spider with legs a, b, c, d, context connecting b to c, and global endpoints a and d. Compute the set of global apparent through-paths meeting P under your formal definition and compare with the right-hand side of Eq. (34). If the two-segment path exists, Eq. (34) must be revised; if it is excluded by a new restriction, that restriction must be stated and then checked against the teleportation and GHZ rewriting movies of Appendix D.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.2.1 defines an apparent through-path as a continuous path that 'crosses no visible tensor-product gap', but no graph-level or category-level formalization of the set L(D) is given. Everything downstream inherits this ambiguity: the boundary fibres L_P(e) in Eq. (20), the local inheritance relation (24), and the local-to-global gluing decomposition (34). The ambiguity is not cosmetic. A multi-port spider is treated as a single vertex whose internal 'through' pairing is never specified; whether a path may enter and leave the same replaced region more than once is never stated. In particular, Eq. (34) implicitly assumes that every global path meeting the replaced subdiagram decomposes uniquely as one external remnant plus one local segment attached to a single boundary-leg pair e. Under the natural graph reading, take P to be one Z-spider with four external legs a, b, c, d, and let the context connect b to c outside, with global endpoints at a and d. A legal simple path enters P at a, leaves at b, follows the external bridge back to c, re-enters P, and exits at d. Its intersection with P is two disjoint local segments (a-b and c-d), so it is not represented by any single term in the disjoint union on the right-hand side of Eq. (34). Thus the gluing principle is either false or missing an additional restriction that is nowhere stated. Since the Definition of a Coecke flow line (Section 3.2.2) is just the existence of a compatible-inheritance chain through such movies, an ill-defined or only visually specified L(D) makes the central claim untestable as written.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper proposes a formalization of Coecke's intuitive notion of \"quantum information flow\" in the language of string-diagram rewriting. Given a semantics-preserving rewriting movie, the authors distinguish apparent through-paths, which are defined for a single frame, from genuine through-paths, which admit compatible inheritance along every rewriting step and terminate in a tensor-decoupled bare-wire factor. A Coecke flow line is then defined as a branch-independent initial through-path that is genuinely through-going in every fixed classical branch. The paper develops a local boundary-fibre analysis of elementary rewrite rules, a local-to-global gluing principle, and a backward-generation algorithm, and illustrates these with the spider-fusion and Hopf rules of the ZX calculus. Applications discussed include quantum teleportation, GHZ-assisted teleportation, entanglement swapping, entanglement distillation, and a tentative connection to holographic bit threads via a companion paper.","tokens_in":39380,"tokens_out":5531,"duration_ms":57926,"significance":"If the main formalization is made precise, the paper would fill a genuine gap: it would turn Coecke's heuristic \"quantum information flow\" into an explicit, testable criterion involving compatible inheritance, rather than a purely visual notion. The paper is deliberately independent of the companion bit-thread paper for its formal core, uses standard ZX-calculus results, and presents worked examples in which the intended flow lines are exhibited frame by frame. The backward-generation algorithm is a useful and concrete contribution. However, the central primitive -- the apparent through-path and the set L(D) of all such paths -- is never given a formal graph-theoretic or categorical definition, and the local-to-global gluing principle as stated is not valid without additional restrictions. These are load-bearing issues for the paper's central claim, although they appear fixable within the manuscript's scope.","major_comments":[{"comment":"The set L(D_k) of apparent through-paths is never formally defined. The phrase \"a continuous path that crosses no visible tensor-product gap\" is a visual predicate: the paper does not specify the underlying graph or spatial model of a string diagram, the allowed behaviour of a path at spiders, cups, and caps, or whether a path may enter and leave the same local region more than once. Since Eq. (20), the local inheritance relation (24), the gluing decomposition (34), and the backward-generation sets C_k in Eq. (47) all quantify over this undefined set, the central formalization is not yet testable. The definition should be replaced by a precise combinatorial characterization, for example paths in a prescribed embedded graph with explicit forbidden transitions at tensor-factor separations, together with stated conventions on repeated vertices and edges.","section":"Section 3.2.1, Definition (Apparent through-path)"},{"comment":"The local-to-global gluing principle is asserted without the uniqueness condition needed for Eq. (34). A global apparent through-path can meet the replaced region in several disjoint segments: take P to be a four-legged Z-spider with boundary legs a, b, c, d, and let the context connect b to c, with global endpoints a and d. A legal path enters P at a, leaves at b, follows the context bridge to c, re-enters P, and exits at d. Its intersection with P is two local segments, so it is not represented by any single term of the form (external remnant, local segment) in the disjoint union on the right-hand side of Eq. (34). Either Eq. (34) must be restricted to paths whose intersection with the replacement region is connected, or the gluing must be defined over sequences of boundary-leg pairs. Without one of these fixes, the global inheritance rule in Eqs. (37)-(39) is not well-defined.","section":"Section 4.1.3, Eqs. (34)-(39)"},{"comment":"The boundary-fibre table for spider fusion leaves the same-side entries as \"1?\", and the accompanying text says that these ambiguities can be dispensed with by decomposing the generalized fusion rule into two elementary fusion rules. However, no boundary-fibre counts for those elementary rules are actually provided. Because the paper's central claim includes the statement that compatible inheritance can be decided locally, rewrite by rewrite, the reader cannot verify the central illustrative calculation. The same-side entries should be computed directly under the formal path definition adopted after fixing the issue raised above, or the elementary-fusion analysis should be carried out in full.","section":"Section 4.3, Table (52)"}],"minor_comments":[{"comment":"There is a typo in the sentence \"We now return to the elementary rewrite rule (17), s ince Pρ and Qρ have the same boundary...\": \"s ince\" should be \"since\".","section":"Section 4.1.2"},{"comment":"The paragraph beginning \"This raises a more general problem concerning string-diagrammatic calculation\" is repeated almost verbatim; one of the two copies should be removed.","section":"Section 3.1"},{"comment":"In the sentence defining C_k, the phrase \"can be certified by the remaining remaining suffix\" contains a duplicated word \"remaining\".","section":"Section 4.2"},{"comment":"The bit-thread correspondence for the HaPPY example is stated to be shown in the companion paper [8] rather than in this manuscript; the text should state explicitly that this comparison is not established as a theorem in the present paper, so that readers do not mistake Fig. 12 for a self-contained derivation.","section":"Section 6.1"}],"recommendation":"major_revision","confidential_remarks":"The paper is best viewed as a conceptual formalization paper rather than a new computational result, and its main claim hinges on making the word \"path\" precise in string diagrams. The companion paper [8] appears to carry the strongest physical claim about bit threads, so the present manuscript should be evaluated on its own formalization. I believe the central content is salvageable, but the undefined primitive and the gluing counterexample need to be addressed before the formalization can be accepted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Good idea, but the core formal object is under-defined. The paper gives a serious attempt to make Coecke's 'quantum information flow' precise, and the distinction between apparent and genuine through-paths is genuinely new. The Coecke flow line definition — a branch-independent apparent through-path that inherits compatibly to a decoupled bare wire in every branch — is a real step forward. The boundary-fibre analysis for ZX rules and the backward-generation algorithm are useful, and the worked examples (teleportation, GHZ swapping, HaPPY) are instructive.\n\nThe soft spot is load-bearing. The set L(D) of apparent through-paths is defined only visually: a continuous path crossing no visible tensor-product gap. There is no graph-level or categorical definition. The local-to-global gluing principle, Eq. (34), asserts each global path through the replaced region decomposes uniquely as a context remnant plus one local segment attached to a boundary-leg pair. That is false under a natural graph reading: a path can enter a four-legged spider at a, leave at b, follow an external bridge to c, re-enter, and exit at d. Its intersection with the replaced region is two disjoint segments, so it does not appear as any single term in the disjoint union. The authors never state a restriction that would rule this out, and they do not prove the decomposition. Since compatible inheritance and Coecke flow lines are defined through L(D), the central formalization is not testable as written.\n\nThis is fixable. Give a precise definition of apparent through-path (e.g., as a path in the underlying graph, modulo some equivalence) and either prove the unique-decomposition property under a stated condition or reformulate the gluing principle. The '1?' entries in the spider-fusion boundary-fibre table should also be resolved at the coarse-grained level or explicitly delegated to the fine-grained rules. Minor issues: duplicated paragraphs in Section 3.1 and the relation to the companion paper [8] is asserted rather than proved.\n\nWho is this for? ZX-calculus and CQM researchers, and people interested in holographic bit threads as physical flows. It deserves a serious referee because the idea is important and the examples are well chosen, but it needs major revision. I would not cite it in this form, but I would bring it to a reading group for discussion.\n\nRecommendation: send to peer review, with a clear message that the formal definition of apparent through-path and the gluing principle must be addressed before acceptance.","headline":"There's a real idea here, but the central formal object — the apparent through-path — is never given a precise definition, and the gluing principle is false or unproven as stated.","tokens_in":39907,"tokens_out":4163,"would_cite":false,"duration_ms":38119,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper turns quantum information flow into an explicit string-diagram object: a through-path that survives every semantics-preserving rewrite until it is a bare-wire factor.","keywords":["quantum information flow","string diagrams","ZX calculus","compatible inheritance","boundary fibres","Coecke flow line","categorical quantum mechanics","bit threads"],"falsifier":"Search for a pair of semantics-preserving rewriting movies that connect the same initial and final diagrams but certify different apparent through-paths as Coecke flows, or a single movie where a path passes every local boundary-fibre check yet dies only when the full context is examined; either would show the definition depends on the chosen witness rather than on the underlying protocol.","tokens_in":38880,"feed_emoji":"🧵","tokens_out":8804,"duration_ms":77440,"temperature":0.7,"pith_summary":"This paper tries to pin down what \"quantum information flow\" means in protocols such as teleportation. It distinguishes an apparent through-path, which merely looks continuous in one diagram, from a genuine through-path, which can be compatibly inherited through every step of a semantics-preserving rewriting movie and ends up as a tensor-decoupled bare wire. It then defines a Coecke flow line as a branch-independent apparent through-path that realizes this genuine through-goingness in every fixed classical branch. If the definition works, a long-standing visual intuition becomes a checkable structural property, measurable locally by counting boundary fibres of elementary rewrite rules.","feed_headline":"Coecke flow line: a path that survives every rewrite to a bare wire","feed_subtitle":"Only branch-independent through-paths that inherit through every semantics-preserving rewrite count as genuine flow.","key_machinery":"The argument rests on four linked pieces. An apparent through-path is a continuous path in a diagram that crosses no visible tensor-product gap. Compatible inheritance is a relation $T_{\\rho}$ between apparent through-paths in consecutive frames of a rewriting movie, inherited through a single rewrite. A boundary fibre $L_P(e)$ collects all local apparent through-paths in a local diagram $P$ meeting the same boundary-leg pair $e$, and the local inheritance relation induced by a rewrite rule $\\rho$ is the disjoint union over $e$ of $L_{P_\\rho}(e) \\times L_{Q_\\rho}(e)$. The local-to-global gluing principle says a global path decomposes into an external remnant and a local segment, so one-step inheritance is determined entirely by the successor boundary fibre; cardinality $0$, $1$, or $>1$ gives death, unique inheritance, or branching. The backward-generation algorithm then starts from the terminal bare wire and pulls examples back through the movie, yielding exactly the certifiable Coecke flow lines.","core_discovery":"The central discovery is a formal definition: a Coecke flow line is an apparent through-path in the initial protocol diagram that, in every fixed branch of the protocol, admits a lineage of representatives inherited step by step through the rewriting movie and whose final representative is carried by a bare-wire factor tensor-decoupled from the rest of the diagram. The paper shows that this definition is testable because inheritance can be decided locally: for each elementary rewrite rule, one compares boundary fibres, the sets of local through-paths joining the same pair of boundary ports, and the cardinality of the successor fibre decides whether a path dies, survives uniquely, or branches. Applied to the ZX calculus, the paper computes these fibres for representative rules, including the Hopf rule, whose right-hand side has an empty fibre and therefore kills inheritance, and spider fusion, which can branch. Physically, a Coecke flow line is read as a constrained, quasi-local, line-like presentation, inside the original strategy diagram, of the target bare-wire morphism factor that the protocol shapes.","pith_inferences":["A natural testable extension is to compile boundary-fibre tables for a complete axiomatization of the ZX calculus; if any elementary rule has a successor fibre of cardinality greater than one in a nontrivial way, the backward algorithm's branching structure becomes a quantitative measure of how many distinct flow-line witnesses exist.","The definition may also give a handle on causality: in protocols where information flows against physical time, as in the original traversal rules, the compatible-inheritance relation makes the direction of a flow a property of the rewriting movie rather than of the underlying process, inviting comparison with causal structure in process theories.","If the bit-thread correspondence is taken seriously, one could use Coecke flow lines to classify non-uniqueness of bit-thread configurations: different rewriting movies may certify different lineages, and their branching fibres might label the family of allowed thread geometries.","The paper leaves the visual criterion \"crosses no visible tensor-product gap\" at the level of diagrams; a natural next step is to formalize it as a graph-theoretic condition on the embedded graph underlying a string diagram."],"forward_implications":["Any protocol whose rewriting movie ends in a tensor-decoupled bare wire now has a certificate: a Coecke flow line exists exactly when the backward-generation algorithm reaches the initial diagram in every branch.","Because inheritance is decided by boundary-fibre cardinalities, a flow line can be checked rule by rule without scanning the whole diagram.","Multiple simultaneous bare-wire factors give multiple Coecke flows, which the paper connects to distilling several Bell pairs from one initial entangled state.","The definition is not tied to the ZX calculus: it applies to any string-diagrammatic calculus with small local rewrite rules and terminal bare-wire factors, and in principle beyond dagger-compact structure.","In holographic tensor-network states, the paper reports that Coecke flow lines obey the same constraints as bit threads, giving bit threads a candidate process-theoretic physical interpretation."],"supporting_citations":[{"why":"Introduces the original quantum-information-flow paths and box-traversal rules that the paper aims to formalize.","marker":"[1]"},{"why":"Gives the concrete and axiomatic account of quantum information flow that supplies the intellectual lineage.","marker":"[2]"},{"why":"Provides the dagger compact closed categorical semantics in which the formalization is framed.","marker":"[5]"},{"why":"Presents the bare-wire and black-line picture of information flow that the terminal-frame condition makes precise.","marker":"[7]"},{"why":"Supplies the ZX calculus used as the concrete rewriting system for boundary-fibre analysis.","marker":"[21-24]"},{"why":"Establishes the completeness of the Clifford ZX rule set, the axiomatic basis for the local inheritance tables.","marker":"[63,64]"},{"why":"Gives the branch-by-branch ZX proof of GHZ-assisted teleportation used as a worked example.","marker":"[34]"}],"fun_headline_variants":["Coecke flow: paths that survive every rewrite to a bare wire","Defining genuine quantum flow via string-diagram rewriting","Inheritance through rewrites: the test for real flow lines","String-diagram rewriting defines robust quantum info flow"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that a usable rewriting system has a reasonably small set of elementary rewrite rules, each supported on a small local diagram with fixed boundary ports; without that, the boundary-fibre inheritance relation and the gluing principle that decide a Coecke flow line simply do not apply.","fun_headline_variants_meta":{"raw":{"variants":["Coecke flow: paths that survive every rewrite to a bare wire","Defining genuine quantum flow via string-diagram rewriting","Inheritance through rewrites: the test for real flow lines","String-diagram rewriting defines robust quantum info flow"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000267,"raw_usage":{"total_tokens":1589,"prompt_tokens":891,"completion_tokens":698,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":507,"completion_tokens_details":{"reasoning_tokens":630}},"tokens_in":507,"tokens_out":698,"duration_ms":6635,"temperature":1.0,"reasoning_tokens":630,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T05:56:38.342689+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Search for a pair of semantics-preserving rewriting movies that connect the same initial and final diagrams but certify different apparent through-paths as Coecke flows, or a single movie where a path passes every local boundary-fibre check yet dies only when the full context is examined; either would show the definition depends on the chosen witness rather than on the underlying protocol.","supporting_citations":[{"cited_title":"The logic of entanglement","cited_arxiv_id":"quant-ph/0402014","evidence_quote":"Introduces the original quantum-information-flow paths and box-traversal rules that the paper aims to formalize."},{"cited_title":"Quantum information-flow, concretely, and axiomatically,","cited_arxiv_id":null,"evidence_quote":"Gives the concrete and axiomatic account of quantum information flow that supplies the intellectual lineage."},{"cited_title":"Quantum Protocols involving Multiparticle Entanglement and their Representations in the zx-calculus,","cited_arxiv_id":null,"evidence_quote":"Gives the branch-by-branch ZX proof of GHZ-assisted teleportation used as a worked example."}],"review_version":1}