{"id":"f3fe93b2-ee53-4046-a090-25b25330f36b","arxiv_id":"2610.28148","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A criterion based on the Müger centralizer of the condensed anyons predicts which point and loop excitations become critical, and a Ginzburg-Landau construction from fusion trees gives field theories for these 2+1D transitions.","lead":"The authors extend two symmetry-based field-theory tools to 2+1D anyon condensation transitions, where non-abelian particles condense and change the topological phase. They also propose a categorical rule predicting which particle and loop excitations become critical at a continuous transition, illustrated on toric code and non-abelian models.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek/deepseek-v4-flash","headline":"Point-excitation rule is ambiguous: Criteria 2 item 3 gives B, Eq. (6) gives B∩B′; the doubled-Ising A2 example contains the split and the paper never says which set predicts criticality.","rationale":"The reader's CONDITIONAL verdict is reasonable, and I do not dispute the internal consistency of most examples. My concern is not primarily the identification K=B′, which is arguably a standard consequence of condensation theory; it is the next inference from K=B′ to the set of critical point anyons. The abstract and Criteria 2 item 3 state the braiding condition literally, which yields the set B by the double-centralizer theorem. The mathematical paragraph then derives a smaller set, B′∩B, by implicitly restricting to endpoints of condensed loops. In the doubled-Ising A2 example these sets differ, and the difference includes σσbar, the very anyon whose condensation drives the transition and whose gap must vanish in a continuous transition. Whichever reading is intended, the central point-excitation criterion is not precisely stated. This is fixable, but it is load-bearing because the paper's advertised novelty is exactly the categorical prediction of critical excitations in non-abelian cases. I therefore keep the reader's conditional verdict rather than accepting the paper as stated.","tokens_in":24183,"tokens_out":24883,"duration_ms":199911,"concrete_test":"Derive the critical point-anyon set for a continuous transition from the local A-module category C_A^loc: compute the image of the induction functor for each x∈B in the doubled-Ising example A2 and check whether the condensed/deconfined/confined split forces the σσbar point excitation to be among the modes whose gap closes, and do the same for ψ1bar and 1ψbar. If the induced module of σσbar is local, or if its gap is required to close by the order-parameter field theory of Section II.B, then the operative prediction is the full set B and Eq. (6) must be revised; if instead only endpoints of condensed loops count, then Criteria 2 item 3 and the abstract need an explicit qualifier that removes σσbar, and possibly the fermionic objects, from the critical set. Either branch settles which formula is the actual claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section III.B, Criteria 2 item 3 states that a point-like anyon becomes critical if it braids trivially with every condensed loop in K. With K equal to the Müger centralizer B′ of B=⟨A⟩, the double-centralizer theorem gives the set of such anyons as (B′)′=B. The immediately following mathematical paragraph instead says that critical point excitations can exist at the end points of condensed loops and are therefore represented by objects in the Müger center Z2(B′)=B′∩B, Eq. (6). These two sets coincide only when B⊆B′. The paper's own doubled-Ising A2 example has B={1,ψψbar,ψ1bar,1ψbar,σσbar} and B′={1,ψψbar}: σσbar∈B∖B′ braids trivially with every loop in K and is the non-abelian condensing anyon whose gap must close in a continuous transition according to the GL construction of Section II.B. Item 3 therefore counts σσbar as critical, while Eq. (6) excludes it. The text never explicitly reconciles whether the intended prediction is B, Z2(B), or the union of endpoint particles and fully condensed anyons; this ambiguity sits exactly in the non-abelian regime advertised in the abstract and in the reader's strongest claim. A reader cannot extract a definite prediction for the doubled-Ising-to-trivial transition without first resolving it.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek/deepseek-v4-flash","summary":"The paper proposes two complementary ways to extract information about 2+1D anyon condensation transitions from the parent braided fusion category C and a condensable algebra A. Section II adapts the authors' previous SymTFT-based Ginzburg-Landau procedure for 1+1D transitions [18] to 2+1D: when all anyons in A have integer quantum dimension, fields are assigned to their internal dimensions, and interactions are generated by fusion trees into the vacuum whose branches all lie in A, with sums over fusion orders for higher-order trees; the construction is worked out for the doubled-Ising algebras A1, A2 and the D(S3) algebras A1, A2, A3, A6. Section III proposes a criterion (Criteria 2) for which excitations become critical at a continuous transition: the loops that remain condensed through the transition form a set K, identified in the mathematical-perspective paragraph with the Müger centralizer B′ of the fusion subcategory B=⟨A⟩; all other loops become critical, with noncontractible-loop splitting scaling as 1/L; and a point anyon becomes critical if it braids trivially with every loop in K, with critical point excitations additionally identified with objects in the Müger center Z2(B′)=B′∩B (Eq. (6)). The authors emphasize that the criterion covers non-abelian condensates and cases with trivial Z2(B), where the 3+1D SymTFT construction of Refs. [15,16] captures no critical point excitations.","tokens_in":24527,"tokens_out":27483,"duration_ms":155921,"significance":"If the criterion is correct, it turns the condensable algebra of the parent order into concrete, falsifiable predictions about which loops and point anyons are critical at a continuous transition, including the 1/L scaling of noncontractible-loop splittings, and it covers non-abelian condensates and the cases where the 3+1D SymTFT construction of Refs. [15,16] has trivial Z2(B) and captures no critical point excitations. The ingredients that are checked are correct: the fusion-tree combinatorics of Sections II.B and II.C, the centralizer computations behind Eq. (6), and the exact toric-code-to-Ising mapping with the universal 1/L splitting used for the toric-code example. The paper is also honest about its scope: the GL rules are presented as proposals, the doubled-Fibonacci and D(S3) candidate models are first order (b343; Refs. [9,15,16]), and the authors call for larger-scale numerics (b333). The weakness is that the two load-bearing steps — the identification of the condensed loop set K with B′, and the point-excitation rule — are asserted heuristically, and the second is stated inconsistently in item 3 versus Eq. (6).","major_comments":[{"comment":"The point-excitation rule is stated as two inequivalent sets, and the discrepancy lands exactly in the non-abelian regime advertised in the abstract. Item 3 (a point anyon is critical iff it braids trivially with every loop in K), combined with K=B′ and the double-centralizer theorem quoted in the text, gives the set (B′)′=B; the mathematical-perspective paragraph instead says critical point excitations are represented by objects in the Müger center Z2(B′)=B′∩B, Eq. (6). The sets differ whenever B⊄B′, which is the doubled-Ising A2 example of Section III.C.3: there B={1,ψψ̄,ψ1̄,1ψ̄,σσ̄} and B′={1,ψψ̄}, so the non-abelian condensing anyon σσ̄ lies in B∖B′, braids trivially with every loop in K, and must have its gap close in a continuous transition by the field theory of Section II.B; item 3 and the sentence 'point-like anyons that become fully condensed ... are critical at the transition' therefore count σσ̄ as critical, while Eq. (6) excludes it. Section III.C.3 mentions only the ψψ̄ point excitation and never states the status of σσ̄, and in the doubled-Fibonacci case (Section III.C.4) K={1} makes item 3 degenerate to 'every point anyon is critical.' The paper must state which set (B, B′∩B, or endpoints of condensed loops together with fully condensed anyons) is the intended prediction and recompute the examples accordingly.","section":"Section III.B (Criteria 2 item 3 and Eq. (6))","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The identification of the physically condensed loop set K with the Müger centralizer B′ is the load-bearing step of the entire criterion, since every loop and point prediction in Criteria 2 and in the examples passes through it, yet it is asserted with heuristic support ('For a more mathematical perspective') rather than derived. The numerical evidence is limited to L=2,3,4 exact diagonalization: at hz>0 the e-loop eigenvalues no longer define exact symmetry sectors (b286), the doubled-Ising test identifies ψψ̄, ψ1̄ and σσ̄ loop content only indirectly through the Q1, Q2 operators and the self-dual toric-code mapping, and the authors themselves note substantial finite-size effects (b332) and state that larger quantum Monte Carlo calculations are needed (b333). Since the doubled-Fibonacci and D(S3) candidate transitions examined in the paper are first order (b343; Refs. [9,15,16]), there is no confirmed continuous non-abelian example checked beyond these small tori; if the physical loop set differs from B′ in that regime, all predictions of Criteria 2 fail.","section":"Section III.B (mathematical-perspective paragraph)","authors_acknowledge":"Larger system calculations, poten- tially using quantum Monte Carlo, are needed to establish the asymptotic scaling. 4.","authors_acknowledge_p":0.8},{"comment":"The completeness and uniqueness claims of the GL construction are asserted rather than established: the rule that independent local Lagrangian terms correspond exactly to fusion trees with all branches in A, and the instruction to sum over fusion orders so that branches remain in A for any order, are justified only by the stated RG-motivated picture, and the real/complex field-assignment rules are validated only in the worked examples. The resulting Lagrangians are checked against known results only in the abelian case (Eq. (1)); for the non-abelian examples (Sections II.B and II.C) the Lagrangians and their consequences (two-step transitions in which one internal dimension condenses first, the expected first-order character of the A3 and A6 transitions, and the claim that the topological term is unimportant) are new, unverified proposals. The authors do present the construction as a proposal, but the paper should state explicitly which outputs of Section II are conjectural, in particular the assertion that the constructed set of Lagrangian terms is complete rather than merely a subset.","section":"Section II.A (fusion-tree construction)","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"The leading term in the torus splitting is written as vLε_m, which grows with L and is inconsistent with the sentence immediately preceding it (the splitting scales as 1/L for a 3D CFT) and with the o(L^{-1}) remainder; presumably v ε_m/L is intended.","section":"Section III.C.1, Eq. (7)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The caption describes the gaps as 'originating from the toric code ground states,' but the figure belongs to the doubled-Ising analysis along the self-dual line hx=hz; the caption appears to have been copied from Fig. 4.","section":"Fig. 5 caption","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The clause that all non-condensed loops 'commute with the condensed loop set' is automatic for contractible loops, as the paper itself notes two sentences later, and is not the operative condition of the criterion; the toric-code discussion in Section III.C.1 relies instead on noncontractible m-loops that anticommute with the condensed e-loop, so the criteria would be clearer if the restriction to contractible loops were stated before item 2 rather than after item 4.","section":"Section III.B, item 2","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The notation Z2(B′C) for the Müger center is used without a definition, and B′ and B′C appear interchangeably; a one-sentence definition would help physics readers.","section":"Section III.B (notation)","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a proposal-style theory paper: both central claims are presented honestly as proposals, and the numerical support is preliminary (L≤4). The main blocker is the internal inconsistency in the point-excitation rule (item 3 vs Eq. (6)), which I judge fixable by a clear statement of the intended prediction. The GL section depends heavily on the authors' companion paper [18] (arXiv:2609.06751); the editor may wish to confirm that this preprint is publicly available in its cited form. The paper is within scope for a field-theory/condensed-matter journal and is competently written. Note that the most novel claim — that the criterion works when Z2(B) is trivial, beyond the SymTFT methods of [15,16] — is exactly the regime (doubled Fibonacci, doubled Ising A2) where the point-excitation rule is most ambiguous, so the revision should prioritize that regime."},"author_rebuttal":null,"desk_editor":{"model":"deepseek/deepseek-v4-flash","letter":"Two things to know up front. The loop part of the criterion is clean: loops that stay condensed through a continuous anyon condensation transition are identified with the Müger centralizer B′ of the subcategory generated by the condensable algebra, all other loops are predicted critical, and the observation that the 3+1D SymTFT constructions of Refs. [15,16] miss the critical loops when Z2(B) is trivial is a genuinely useful gap to have on record. The point-excitation rule is not clean, and the ambiguity sits exactly in the non-abelian regime the abstract advertises.\n\nWhat is actually new and what the paper does well. The GL construction is a transparent adaptation of the authors' own 1+1D procedure, scoped honestly (integer quantum dimensions for the condensing anyons), with two substantial worked examples, doubled Ising A2 and D(S3) A6. The fusion-tree combinatorics checks out; the toric-code-to-Ising mapping at hz = 0 is exact and Eq. (7) is right; the D(ZN), Potts, and first-order Fibonacci/D(S3) checks all agree with known results. The authors flag the first-order cases instead of using them as support and state their own finite-size caveats. This is a well-scoped proposal, not a dressed-up claim.\n\nThe soft spot is real. Criteria 2, item 3 says a point-like anyon becomes critical if it braids trivially with every condensed loop. With K = B′, the double-centralizer theorem gives exactly B, the subcategory generated by the condensable algebra. The paragraph immediately after instead puts critical point excitations in Z2(B′) = B∩B′, Eq. (6). The two sets coincide only when B ⊆ B′. In the doubled-Ising A2 example, B = {1, ψψbar, ψ1bar, 1ψbar, σσbar} and B′ = {1, ψψbar}; σσbar braids trivially with every condensed loop, so item 3 counts it critical while Eq. (6) excludes it. The application section mentions only the ψψbar point and never says which rule it is applying. This matters physically: in the authors' own GL theory of Section II.B, the mass term of the condensing σσbar component is what drives the second step of the transition, so excluding σσbar from the critical point excitations looks wrong, not merely under-specified. The text never reconciles the two statements, and a reader cannot extract a definite point-spectrum prediction for the doubled-Ising transition without first resolving the question.\n\nThe secondary concerns are the ones the reader already weighed: the identification of the condensed loops with B′ is heuristic rather than derived, and the numerical support is exact diagonalization on L = 2, 3, 4 tori, with the authors' own caveats about finite-size effects and about e-loop expectation values no longer being exact quantum numbers at hz > 0. Both are acknowledged in the text; they bound the paper as a proposal rather than an established result.\n\nWho this is for: people working on anyon condensation and SymTFT who want an organizing tool and a list of concrete consistency checks. It deserves a serious referee. My recommendation is to send it to review, with the reconciliation of the point-excitation rule as the required revision.","headline":"The loop rule is clean and the examples check out, but the point-excitation rule is ambiguous exactly where the paper needs it: item 3 of Criteria 2 predicts B, Eq. (6) predicts B∩B′, and the doubled-Ising application never confronts the difference.","tokens_in":25139,"tokens_out":20688,"would_cite":true,"duration_ms":141234,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"pith_extraction":{"msc":[],"pacs":["05.30.Pr","64.70.Tg"],"model":"deepseek/deepseek-v4-flash","headline":"This paper argues that the categorical data of the parent order and its condensable algebra alone determine which point and loop excitations turn critical at anyon condensation transitions.","keywords":["anyon condensation","topological order","condensable algebra","Müger centralizer","critical excitations","non-abelian anyons","Ginzburg-Landau field theory","torus spectrum"],"falsifier":"At the toric-code-to-trivial transition with $0<h_z<h_x$, measure the splitting $\\Delta_m(L)$ induced by a noncontractible $m$ loop on a finite torus; the criterion predicts $\\Delta_m(L)\\sim 1/L$ at the critical field $h_{x,c}\\approx 0.33$, so a Monte Carlo or tensor-network computation at system sizes beyond $L=4$ that finds an exponentially small or finite splitting would show the condensed-loop set is not $B'=\\{1,e\\}$ as claimed. More generally, a continuous condensation transition in which a loop outside $B'$ remains condensed, or in which a point excitation braiding trivially with every loop in $B'$ stays gapped, would refute the rule.","tokens_in":25481,"feed_emoji":"🌀","tokens_out":17221,"duration_ms":103546,"temperature":0.7,"pith_summary":"Anyon condensation transitions—in which a collection of bosonic quasiparticles condenses and reshuffles the topological order—are notoriously hard to analyze, and hardest when the condensing anyons are non-abelian. This paper argues that the categorical data of the parent topological order and of the condensable algebra describing the condensate are enough to predict the universal critical behaviour: which loop-like and point-like excitations become gapless at a continuous transition, and what kind of field theory governs it. Two tools are developed: a Ginzburg-Landau construction that writes down Lagrangians from fusion trees of the condensing anyons (valid when those anyons have integer quantum dimensions), and a criterion stating that the anyon loops that remain condensed through the transition form the Müger centralizer of the subcategory generated by the condensate, while every other loop, and every point excitation braiding trivially with all condensed loops, becomes critical. If correct, the criterion would let one read critical excitations off the data of the neighbouring gapped phases alone, including at non-abelian condensations where the standard 3+1D symmetry-TFT construction sees nothing.","feed_headline":"Predict which anyons go critical from the condensate alone","feed_subtitle":"New criterion reads critical excitations off the condensate, covering non-abelian anyons other methods miss.","key_machinery":"The load-bearing objects are the condensable algebra $\\mathcal{A}$ (the set of anyons that condense together along with their fusion channels), the fusion subcategory $B=\\langle \\mathcal{A}\\rangle$ that it generates, and the Müger centralizer $B'$—the anyons with trivial braiding against every object of $B$. The centralizer does the work: the paper proposes that the condensed loop set $K$ is exactly $B'$, so on the torus the ground states differing by inserting a loop from $B'$ stay exponentially close throughout, while a loop outside $B'$ produces a splitting that vanishes as $1/L$ at the critical point. Point excitations then become critical precisely when they lie in $Z_2(B')=Z_2(B)$, the intersection of the centralizer with the generating subcategory. For the field theory construction, the machinery is the fusion-tree expansion: each Lagrangian term traces a tree of condensing anyons fusing to the vacuum, summed over F-moves so no branch ever leaves $\\mathcal{A}$, with real or complex fields attached to the internal dimensions of each non-abelian anyon according to its self-statistics, and a topological Chern-Simons term encoding the mutual non-locality of different field components.","core_discovery":"On the paper's own terms: at a direct continuous anyon condensation transition out of a topological order $\\mathcal{C}$ driven by a condensable algebra $\\mathcal{A}$, let $B=\\langle \\mathcal{A}\\rangle$ be the fusion subcategory generated by the condensing anyons. Loops of anyons that split into only deconfined or condensed components remain condensed across the transition and form the set $K$, identified with the Müger centralizer $B'$ (the anyons braiding trivially with everything in $B$); inserting such a loop along a noncontractible torus cycle costs at most an exponentially small splitting. Every other anyon loop becomes critical, meaning the splitting from inserting it vanishes algebraically with system size, as $1/L$ for a 3D CFT. A point-like anyon becomes critical exactly if it braids trivially with every condensed loop in $K$; otherwise it stays gapped. Because $\\mathcal{C}$ is modular, the double centralizer theorem gives $(B')'=B$, so the critical point excitations sit in the Müger center $Z_2(B)=B'\\cap B$—the same object that earlier 3+1D SymTFT constructions use—while the new loop criterion additionally covers the case $Z_2(B)=\\{1\\}$, where that bulk construction is trivial and misses the critical loops entirely. The paper also constructs Ginzburg-Landau Lagrangians for these transitions: when the condensing anyons have integer quantum dimensions, fields are assigned to the internal dimensions of each condensing anyon, and each allowed interaction term corresponds to a fusion tree into the vacuum whose branches always stay inside $\\mathcal{A}$.","pith_inferences":["Editorial extension: a fusion 2-category treatment that includes descendant line defects such as Cheshire strings—explicitly left out by the paper—might turn the heuristic identification of the condensed loops with the Müger centralizer into a theorem, or reveal corrections to which loops count as condensed.","Editorial extension: the same logic suggests a general ordering principle for 2+1D critical points—gaplessness is carried first by line-like objects, and point-like criticality is derivative, arising only where points braid trivially with the surviving line condensate—so future field theories of such transitions should start from the 2-category structure of line defects rather than from point orde","Editorial extension: the paper's own numerical evidence stops at $L=4$, where loop eigenvalues are no longer exact quantum numbers; a quantum Monte Carlo computation of the $m$-loop splitting at the toric code transition beyond $L=4$ would settle whether the centralizer identification survives in the thermodynamic limit.","Editorial extension: because the criterion is conditional on continuity, it doubles as a first-order-detection device—a proposed lattice realization whose predicted critical loops stay gapped (as happens for doubled Fibonacci) is evidence the transition is actually first order rather than evidence against the criterion."],"forward_implications":["At any continuous anyon condensation transition, which excitations turn critical is fixed by the condensable algebra and parent category alone: loops outside the Müger centralizer $B'$ become critical, and a point excitation becomes critical iff it braids trivially with every loop in $B'$.","The criterion reaches cases where the 3+1D SymTFT construction of earlier work is blind: when the Müger center $Z_2(B)$ is trivial the SymTFT bulk is trivial and misses the critical loop excitations, which the loop criterion still predicts.","For the toric code in a magnetic field the criterion predicts that the $e$ point and the $m$ and $f$ loops turn critical, and exact diagonalization on $L=2,3,4$ tori shows the relevant level crossings moving toward $h_{x,c}\\approx 0.33$ with the expected $1/L$ scaling.","For the doubled Ising and $S_3$ quantum double orders the construction yields explicit Lagrangians (Ising-type and $Z_3$ clock-type terms) and, via ungauging relations, concrete numerical signatures such as the order parameter odd under the $e$-$m$ exchange at the self-dual toric code transition."],"supporting_citations":[{"why":"provides the condensable-algebra description of condensation (splitting, identification, confinement) on which the criterion's language of condensed, deconfined, and confined components relies.","marker":"[1]"},{"why":"defines anyon condensation through tensor categories and the local A-module category of deconfined anyons, used to state which split components stay deconfined after the transition.","marker":"[2]"},{"why":"the 1+1D Ginzburg-Landau procedure in the SymTFT setting that Section II adapts to 2+1D condensation transitions.","marker":"[18]"},{"why":"the SymTFT criterion for which topological defects become gapless at 1+1D transitions, whose structure Criteria 2 generalizes to 2+1D loops and points.","marker":"[19]"},{"why":"constructs a 3+1D SymTFT bulk from the Müger center for non-abelian proliferation transitions; the comparison target for cases where the present criterion sees critical loops the bulk misses.","marker":"[15]"},{"why":"the companion construction using $Z_2(B)$ for the same transitions, which the paper's loop criterion goes beyond when $Z_2(B)$ is trivial.","marker":"[16]"},{"why":"supplies the double centralizer theorem for modular categories, used to show $Z_2(B')=Z_2(B)$ for the critical point excitations.","marker":"[57]"},{"why":"gives the finite-size torus spectrum formalism with universal $1/L$ splittings used to diagnose critical loop excitations numerically.","marker":"[61]"},{"why":"the abelian doubled Chern-Simons description of self-dual toric code transitions, structurally similar to the topological term in the constructed Lagrangians and used to connect the doubled Ising condensation to the self-dual toric code transition.","marker":"[12]"},{"why":"the tensor-network study that finds the doubled Fibonacci string-net transition first order, cited as the reason that model cannot test the criterion for continuous transitions.","marker":"[8]"}],"fun_headline_variants":["Anyon loops tell you which excitations go critical","Condensate alone predicts critical anyons, including non-abelian","Loop criterion finds critical excitations at anyon condensation","Critical anyons read off the condensate via loop braiding","Condensate's centralizer marks the anyons that go critical"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The load-bearing premise is that the loops remaining condensed through the transition are exactly the Müger centralizer $B'$ of the subcategory generated by the condensable algebra, an identification supported by heuristic argument and by exact diagonalization on only $L=2,3,4$ tori rather than derived from the microscopic dynamics, so if the true condensed-loop set differs the predicted critical loops and points fail.","fun_headline_variants_meta":{"raw":{"variants":["Anyon loops tell you which excitations go critical","Condensate alone predicts critical anyons, including non-abelian","Loop criterion finds critical excitations at anyon condensation","Critical anyons read off the condensate via loop braiding","Condensate's centralizer marks the anyons that go critical"]},"model":"deepseek/deepseek-v4-flash","effort":"low","cost_usd":0.000163,"raw_usage":{"cost":0.000162799,"is_byok":false,"cost_details":{"upstream_inference_cost":0.000162799,"upstream_inference_prompt_cost":3.1031e-05,"upstream_inference_completions_cost":0.000131768},"total_tokens":1845,"prompt_tokens":1121,"completion_tokens":724,"prompt_tokens_details":{"audio_tokens":0,"video_tokens":0,"cached_tokens":975,"cache_write_tokens":0},"completion_tokens_details":{"audio_tokens":0,"image_tokens":0,"reasoning_tokens":652}},"tokens_in":1121,"tokens_out":724,"cost_source":"openrouter_usage","duration_ms":8215,"image_pages":[],"temperature":1.0,"billed_attempts":1,"reasoning_tokens":652,"transport_provider":"openrouter","cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-10-21T08:44:26.370937+00:00","model_set":{"reader":"deepseek/deepseek-v4-flash"},"falsifier":"At the toric-code-to-trivial transition with $0<h_z<h_x$, measure the splitting $\\Delta_m(L)$ induced by a noncontractible $m$ loop on a finite torus; the criterion predicts $\\Delta_m(L)\\sim 1/L$ at the critical field $h_{x,c}\\approx 0.33$, so a Monte Carlo or tensor-network computation at system sizes beyond $L=4$ that finds an exponentially small or finite splitting would show the condensed-loop set is not $B'=\\{1,e\\}$ as claimed. More generally, a continuous condensation transition in which a loop outside $B'$ remains condensed, or in which a point excitation braiding trivially with every loop in $B'$ stays gapped, would refute the rule.","supporting_citations":[{"cited_title":"Condensate-induced transitions between topologically ordered phases, Phys. Rev. B79, 045316","cited_arxiv_id":null,"evidence_quote":"provides the condensable-algebra description of condensation (splitting, identification, confinement) on which the criterion's language of condensed, deconfined, and confined components relies."},{"cited_title":"Anyon condensation and tensor categories, Nuclear Physics B886, 436","cited_arxiv_id":null,"evidence_quote":"defines anyon condensation through tensor categories and the local A-module category of deconfined anyons, used to state which split components stay deconfined after the transition."},{"cited_title":"Ginzburg-Landau Theory for Non-Invertible Symmetry-Breaking Transitions","cited_arxiv_id":"2609.06751","evidence_quote":"the 1+1D Ginzburg-Landau procedure in the SymTFT setting that Section II adapts to 2+1D condensation transitions."},{"cited_title":"S. Sch¨ afer-Nameki, and A. Warman, Hasse diagrams for gapless spt and ssb phases with non-invertible symmetries, SciPost Phys.19, 113","cited_arxiv_id":null,"evidence_quote":"the SymTFT criterion for which topological defects become gapless at 1+1D transitions, whose structure Criteria 2 generalizes to 2+1D loops and points."},{"cited_title":"Proliferation Transitions for Non-Abelian Anyons","cited_arxiv_id":"2608.12303","evidence_quote":"constructs a 3+1D SymTFT bulk from the Müger center for non-abelian proliferation transitions; the comparison target for cases where the present criterion sees critical loops the bulk misses."},{"cited_title":"Note on proliferation transitions of non-Abelian anyons","cited_arxiv_id":"2608.20604","evidence_quote":"the companion construction using $Z_2(B)$ for the same transitions, which the paper's loop criterion goes beyond when $Z_2(B)$ is trivial."},{"cited_title":"M¨ uger, On the structure of modular categories, Proceedings of the London Mathematical Society87, 291 (2003)","cited_arxiv_id":null,"evidence_quote":"supplies the double centralizer theorem for modular categories, used to show $Z_2(B')=Z_2(B)$ for the critical point excitations."},{"cited_title":"L.-P. Henry, S. Sachdev, and A. M. L¨ auchli, Universal signatures of quantum crit- ical points from finite-size torus spectra: A window into the operator content of higher-dimensional con- formal field theories, Phys. Rev. Lett.117, 210401","cited_arxiv_id":null,"evidence_quote":"gives the finite-size torus spectrum formalism with universal $1/L$ splittings used to diagnose critical loop excitations numerically."},{"cited_title":"Self-dual Higgs transitions: Toric code and beyond","cited_arxiv_id":"2601.20945","evidence_quote":"the abelian doubled Chern-Simons description of self-dual toric code transitions, structurally similar to the topological term in the constructed Lagrangians and used to connect the doubled Ising condensation to the self-dual toric code transition."},{"cited_title":"Tensor-network approach to phase transitions in string-net models, Phys. Rev. B100, 245125","cited_arxiv_id":null,"evidence_quote":"the tensor-network study that finds the doubled Fibonacci string-net transition first order, cited as the reason that model cannot test the criterion for continuous transitions."}],"review_version":1}