{"id":"b3d88c17-26c4-46e0-8e0d-ab0da88874d2","arxiv_id":"1908.08049","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"New bulk commutation diagnostics coarsely sort translation invariant 3D stabilizer codes into TQFT, foliated type-I, fractal type-I, or type-II topological order.","lead":"Physicists introduce bulk commutation matrices that sort three-dimensional topological stabilizer models into four qualitative phases by measuring how particles can move. The tools are applied to Haah's 18 cubic codes and other fracton models, producing a reference zoo that separates TQFT, foliated type-I, fractal type-I, and type-II orders.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Universal sorting claim is not certified: finite-width and finite-order checks leave the type-II and fractal-type-I labels in Table III open to large-width string operators.","rationale":"The reader's CONDITIONAL verdict is appropriate. The paper introduces genuinely useful diagnostics and is unusually explicit about its limitations, including the absence of a string-width bound. However, the strongest claim is a universal sorting claim: that the procedure sorts any translation-invariant 3D topological stabilizer model into TQFT, foliated type-I, fractal type-I, or type-II. That claim requires either a proof that finite diagnostic shapes and finite-order constraints suffice, or an upper bound on the relevant width and order. Neither is provided. The finite-width and finite-order loophole is therefore load-bearing: it directly threatens the type-II assignments in Table III, since a string operator wider than the width-3 rods would flip such a label. The HH-II inconclusive row demonstrates that the procedure can fail to decide; the same mechanism could silently corrupt rows marked as conclusive. This concern does not warrant rejection, because the paper's stated aim of providing practical tools is met and the explicit limitations are disclosed. But the universal phrasing of the central claim should be read as conditional, matching the reader's verdict. No change to the verdict is needed.","tokens_in":37907,"tokens_out":6303,"duration_ms":71373,"concrete_test":"For a representative type-II candidate, e.g., cubic code CC2, perform an exhaustive algebraic search for string operators of width w = 4, 5, ..., 20 using the Laurent-polynomial module method of Haah, and independently re-run the Appendix A cleaning conditions for all Fig. 7 pair-creation configurations at fourth and fifth order. If any nonzero flat-rod commutation rank or any non-deformable configuration appears, the Table III type-II assignment is false. To settle the general claim, prove or disprove an upper bound on the required string-operator width and cleaning order for translation-invariant stabilizer models; without such a bound, no finite-width computation can certify the universal sorting statement.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The sorting procedure in Sec. III A treats zero flat-rod commutation ranks plus full deformability as sufficient to certify type-II, and uses the scaling of membrane-membrane ranks to distinguish foliated from fractal type-I. Both steps are verified only at finite scale: flat-rods are taken three stabilizer generators wide, membranes two generators wide, and deformability checks stop at third-order constraints, as stated in the Table III caption and Sec. III B. The authors explicitly concede, 'It would be interesting if one could upper bound the string operator width that needs to be considered for translation invariant topological stabilizer models to close this loophole, but we do not have such a bound presently.' Because a rigid string operator wider than the checked rods, or one requiring higher-order cleaning constraints, would change a type-II label to fractal type-I and could alter other assignments, the central claim that the tools sort any translation-invariant 3D topological stabilizer model is not yet established. The HH-II '?' row is the visible manifestation of this gap; the same gap is silent for rows marked '✓'.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a set of bulk diagnostics for translation-invariant topological stabilizer models in three dimensions: flat-rod string-membrane commutation matrices, membrane-membrane commutation matrices, a sufficient-condition cleaning lemma for deforming pair-creation operators onto flat rods, and generalized Gauss's-law mobility tests. These tools are applied to a zoo of models, including Haah's cubic codes, X-cube, checkerboard, Chamon, Sierpinski fractal spin liquid, and the Hsieh-Halasz models, leading to a proposed coarse sorting into TQFT, foliated type-I, fractal type-I, and type-II phases (Tables I-III). The authors also state conjectures relating independent 3D particles to 3D toric-code copies and 2D particles to 2D toric-code stacks.","tokens_in":38190,"tokens_out":5219,"duration_ms":53444,"significance":"If the universal sorting claim holds, this is an important step toward a practical classification of 3D stabilizer topological order, extending the 2D S-matrix program into the fracton regime. The paper contains real technical content: the local cleaning lemma in Appendix A is a general sufficient condition for deformability, and Appendix B gives a concrete, non-obvious deformation analysis for cubic code 8. The diagnostics are defined directly from stabilizer generators and are not fitted to the classification; they also avoid boundary-condition and entanglement-entropy artifacts. The main risk is that the central claim is stronger than what the finite checks establish, and the authors candidly acknowledge this. The paper is therefore valuable as a systematic toolbox and a conjecture-rich framework, but its headline sorting claim is conditional on unproven finite-size and finite-order sufficiency assumptions.","major_comments":[{"comment":"The sorting procedure certifies type-II assignments from zero flat-rod commutation ranks together with full deformability, but in practice deformability is checked only up to third-order constraints and with flat rods three stabilizer generators wide and membranes two generators wide. The authors explicitly state that they have no upper bound on the required string-operator width. A rigid string operator wider than the checked rods, or one requiring higher-order cleaning constraints, would move a model marked type-II into fractal type-I. The HH-II '?' row is the visible manifestation of this gap, but the same gap is silent for rows marked '✓'; consequently the universal claim in §IV that the tools 'sort any translation invariant topological stabilizer model' is not yet established. This does not invalidate the diagnostics, but the scope of the claim should be qualified accordingly.","section":"§III B and Table III caption"},{"comment":"There is a direct inconsistency in the deformability flag for cubic code 0. Table III lists CC0 as '×' (not fully deformable), consistent with the main-text statement that its only rigid string operators run along non-lattice directions, while the operator data table in Appendix D 2 a (Table XX) lists CC0 as '✓'. Since the deformability flag is one of the two inputs separating type-II from fractal type-I, this contradiction must be resolved; as printed, the two tables give opposite evidence for the same model.","section":"Table III vs Appendix D 2 a"}],"minor_comments":[{"comment":"The phrase 'We compliment the string-membrane configurations' should read 'complement'.","section":"§II B"},{"comment":"The sentence 'Over the coarse of this example centric study' contains typos: 'coarse' should be 'course', and 'example centric' is needlessly informal.","section":"§IV"},{"comment":"The heading 'This is a fracotn model' contains a typo ('fracton'), and the same subsection repeats 'This model has has'.","section":"Appendix D 3 d"},{"comment":"The placeholders n1 and n2 are only described as nonzero width-dependent numbers; readers would benefit from explicit values or formulas for at least one representative model, since the numerical ranks themselves are a key output of the paper.","section":"Table III caption"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid contributions-style manuscript whose main tools are sound as far as they go, but the headline claim of a complete sorting procedure overreaches the demonstrated finite checks. I recommend major revision rather than rejection: the authors should either supply a proof that finite-order constraints and the finite rod/membrane widths suffice for the classes considered, or substantially soften the universal sorting claim and present the method as a heuristic diagnostic with explicit caveats. The CC0 table inconsistency should also be fixed before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe useful thing here is the toolbox, not the universal claim. Dua, Kim, Cheng, and Williamson generalize the 2D S-matrix idea to 3D by looking at flat-rod string-membrane and membrane-membrane commutation ranks plus deformability of pair-creation operators. The construction is non-routine: the naive string-membrane commutation matrix fails for rigid strings, and the membrane-membrane configuration is new. The cleaning lemma in Appendix A is a real, generally applicable algorithmic criterion, and Appendix B runs through CC8 in enough detail to be checked by hand. That alone is worth having.\n\nThe paper also earns its keep as a reference. Table III sorts a large model zoo including all 18 cubic codes, X-cube, checkerboard, Chamon, SFSL, HH-I/HH-II. New assignments like CC7/8/10 as type-II and CC11-17 as fractal type-I with planons are concrete, useful claims that others can test.\n\nThe soft spot is what the stress test flagged, and it is real but not concealed. The type-II labels rely on zero commutation ranks for rods only three generators wide and deformability checks that stop at third order. The authors explicitly say there is no bound on the string width that must be considered. So the central 'sort any model' statement is a working conjecture, not a theorem. I agree with their own framing: the HH-II row with a '?' is the visible symptom, and the same limitation is silent in the checked rows. That keeps me from calling the universal sorting claim established. It does not undermine the value of the diagnostics or the specific assignments as strong numerical and analytic evidence.\n\nNo code or machine-readable data accompanies the paper, and the membrane-membrane scaling is reported only for square aspect ratio with a note about diagonal planons. Minor.\n\nWho is this for? People working on fracton classification and stabilizer model phenomenology will want it on the shelf. It is a solid craft paper with honest limitations. A serious referee should engage with it; the main asks are to separate certified results from conjectures more sharply and to publish the numerics.\n\nMy recommendation: send it to peer review. It will be a useful reference even if the conjectures take years to settle.","headline":"Genuinely useful coarse-graining diagnostics for 3D stabilizer models, with the main caveat honestly stated: the sorting is certified only up to finite width and finite order.","tokens_in":38655,"tokens_out":1430,"would_cite":true,"duration_ms":15340,"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":"Bulk commutation quantities sort 3D topological stabilizer models into four phase classes.","keywords":["topological stabilizer models","fracton topological order","type-II fracton phases","string operators","membrane operators","commutation matrix invariant","flat-rod configurations","three-dimensional topological phases"],"falsifier":"Compute the deformability constraints beyond third order for the model labeled HH-II and search for string operators wider than three stabilizer generators: if a nontrivial string operator is found, the model is fractal type-I despite the paper’s inconclusive type-II-consistent data, and the finite-order flat-rod procedure is shown to be incomplete.","tokens_in":37681,"feed_emoji":"🧩","tokens_out":10516,"duration_ms":89952,"temperature":0.7,"pith_summary":"This paper claims that three bulk diagnostics—flat-rod string-membrane commutation matrices, membrane-membrane commutation scaling, and deformability of pair-creation operators—are sufficient to sort any translation-invariant topological stabilizer model in three dimensions into one of four qualitatively distinct classes: TQFT, foliated type-I, fractal type-I, and type-II. The result matters because three-dimensional fracton order has so far resisted the completeness that holds in two dimensions, where the S-matrix invariant fully classifies the phases. The diagnostics are based purely on bulk operators, so they are insensitive to boundary conditions and avoid the spurious contributions that plague entanglement-entropy measures. Applying the procedure yields concrete assignments for the cubic codes, the X-cube, the checkerboard, and the fractal spin liquid, among others.","feed_headline":"Sorting 3D topological stabilizer models into four phase types","feed_subtitle":"Bulk commutation matrices and deformability tests classify TQFT, foliated, fractal, and type-II orders.","key_machinery":"The central objects are the flat-rod commutation matrices $C_{i,j}$, defined over $\\mathbb{Z}_2$ by whether a string operator $S^r_i$ on a flat-rod configuration anti-commutes with a membrane operator $S^m_j$; their rank counts independent anti-commuting string-membrane pairs. These are supplemented by the membrane-membrane commutation matrix, whose rank scaling with membrane size separates foliated from fractal type-I, and by the local-cleaning condition $\\ker(C_{\\mathrm{out}}(a,A)\\Omega)=\\operatorname{Im}(C_{\\mathrm{in}}(a,A))$, which certifies when a pair-creation operator can be deformed to a flat-rod configuration. The intersection of generalized Gauss’s laws determines the minimal mobility dimension of excitations. Together these are the machinery claimed to be sufficient for the sorting.","core_discovery":"The central claim is that the $\\mathbb{Z}_2$-rank of the flat-rod commutation matrix, the scaling of the membrane-membrane commutation rank, and a local-cleaning test of deformability together give an if-and-only-if discrimination of the four classes. For a model whose pair-creation operators are all deformable to flat rods, equal non-zero flat-rod ranks across the 3D, 2D, and 1D configurations characterize TQFT or a stack of TQFT with type-II, while all ranks zero characterize type-II, and unequal ranks characterize type-I with rigid string operators. For non-deformable models, the membrane-membrane rank scales linearly for foliated type-I and stays constant or linear with fluctuating corrections for fractal type-I. The authors implement the sorting on a zoo of known stabilizer models, concluding that cubic codes 1–4, 7, 8, and 10 are type-II, that the remaining cubic codes are fractal type-I, and that the model labeled HH-II remains inconclusive.","pith_inferences":["The four-class partition is a coarse sort, not a full phase classification; the paper leaves open whether rigid strings confined to non-lattice directions or widths beyond the numerical limits would break the flat-rod completeness, so the type labels should be read as conditional on those finite checks.","A natural test of the conjecture that $n^{3D}_{\\mathrm{rods}}$ counts disentangled 3D toric code copies is to take an unknown model, stack it with a known 3D toric code, and check that the 3D flat-rod rank increases by one; the paper does not perform this additivity test.","The membrane-membrane scaling distinction between ‘constant’ and ‘linear with fluctuations’ is qualitative; measuring the variance of the rank over system size would give a quantitative discriminator and could be applied beyond the square membranes with aspect ratio 1 considered here.","If the sorting is phase-relevant, the diagnostics should be stable under local unitary circuits and under stacking with trivial systems; checking invariance under a known locality-preserving Clifford circuit, such as the one that maps cubic code 16 to cubic code 15, would test whether the invariants are truly invariant rather than model-dependent."],"forward_implications":["If the sorting is correct, the known 3D stabilizer models split cleanly: the X-cube, checkerboard, and HH-I models are foliated type-I; the fractal spin liquid and most cubic codes are fractal type-I; cubic codes 1–4, 7, 8, and 10 are type-II; and the 3D toric code is TQFT.","The value of the 3D flat-rod commutation rank conjecturally counts the number of copies of 3D toric code that can be disentangled from a stabilizer Hamiltonian, giving a quantitative invariant rather than just a class label.","A model with fully deformable pair-creation operators and all flat-rod ranks zero has no string operators and is type-II, while full deformability together with equal non-zero ranks is an ‘if and only if’ condition for TQFT or a stack of TQFT plus type-II up to local unitary.","The bulk-only nature of the diagnostics avoids dependence on boundary conditions and spurious topological entanglement entropy, so the same tools can be applied to large or coarse-grained systems without finite-size ambiguity.","The conjectures imply that in an entanglement renormalization flow, stabilizer models with only 3D particles flow to fixed points, while each independent 2D particle yields a stack of 2D toric codes that can be extracted in the flow."],"supporting_citations":[{"why":"Establishes that 2D translation-invariant topological stabilizer models are fully classified by copies of toric code, the completeness result this work seeks to extend.","marker":"[4]"},{"why":"Provides the algebraic framework for 2D stabilizer classification and the notion of local unitary equivalence used throughout.","marker":"[5]"},{"why":"Introduces the cubic codes and the no-strings proof whose flat-rod deformation ideas motivate the 48 flat-rod configurations.","marker":"[13]"},{"why":"Supplies the Hamiltonian forms of cubic codes 1–10 used in the sorting data of Table III.","marker":"[18]"},{"why":"Defines the X-cube and checkerboard models and the type-I/type-II fracton nomenclature the paper refines into foliated and fractal type-I.","marker":"[23]"},{"why":"Provides the models labeled HH-I and HH-II whose classification (foliated type-I and inconclusive, respectively) is one of the paper’s results.","marker":"[31]"},{"why":"Introduces generalized Gauss’s laws and materialized symmetries used to determine the minimum mobility dimension of excitations.","marker":"[43]"},{"why":"Defines foliated fracton order, the structural notion distinguishing foliated from fractal type-I models.","marker":"[51]"},{"why":"Introduces the local-unitary-invariant S-matrix and commutation matrix ideas that the flat-rod and membrane-membrane matrices generalize.","marker":"[66]"}],"fun_headline_variants":["Four-way split for 3D topological stabilizer models","Commutation invariants sort 3D stabilizer phases","3D stabilizer models sorted into four phases","Sorting 3D stabilizer phases: TQFT, foliated, fractal, type-II","Bulk commutation ranks sort 3D stabilizer phases"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The sorting procedure assumes that any string operator deformable in three dimensions can be deformed onto one of the finite set of flat-rod shapes aligned with lattice directions, and that full deformability can be certified from finitely many local commutation constraints; the paper does not bound the required string width or the order of constraints, so a model with a rigid string only in an unconsidered shape or at a width beyond the numerics would be assigned to the wrong class.","fun_headline_variants_meta":{"raw":{"variants":["Four-way split for 3D topological stabilizer models","Commutation invariants sort 3D stabilizer phases","3D stabilizer models sorted into four phases","Sorting 3D stabilizer phases: TQFT, foliated, fractal, type-II","Bulk commutation ranks sort 3D stabilizer phases"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001416,"raw_usage":{"total_tokens":5672,"prompt_tokens":854,"completion_tokens":4818,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":470,"completion_tokens_details":{"reasoning_tokens":4729}},"tokens_in":470,"tokens_out":4818,"duration_ms":30773,"temperature":1.0,"reasoning_tokens":4729,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:51:10.741932+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the deformability constraints beyond third order for the model labeled HH-II and search for string operators wider than three stabilizer generators: if a nontrivial string operator is found, the model is fractal type-I despite the paper’s inconclusive type-II-consistent data, and the finite-order flat-rod procedure is shown to be incomplete.","supporting_citations":[{"cited_title":"Fractons from Partons","cited_arxiv_id":"1703.02973","evidence_quote":"Provides the models labeled HH-I and HH-II whose classification (foliated type-I and inconclusive, respectively) is one of the paper’s results."},{"cited_title":"An invariant of topologically ordered states under local unitary transformations","cited_arxiv_id":"1407.2926","evidence_quote":"Introduces the local-unitary-invariant S-matrix and commutation matrix ideas that the flat-rod and membrane-membrane matrices generalize."}],"review_version":1}