{"id":"489d53c1-2a33-47da-ac8f-76a5ab8418b9","arxiv_id":"2504.18501","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The authors classify, via Galois theory of cyclotomic polynomials, when finite duality symmetry groups of 4d BF-type SymTFTs admit invariant Lagrangian boundary conditions, thereby determining when symmetry-preserving gapped or gapless IR phases exist.","lead":"This paper gives a general rule for when certain non-invertible symmetries in four-dimensional quantum field theory can survive in a gapped phase, and when they force the theory to stay gapless or break spontaneous symmetry. The rule is a simple condition on prime numbers and cyclotomic polynomials that unifies many previously separate examples.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The no-go inference from 'no invariant Lagrangian' to 'no trivially gapped phase' rests on the coarse SymTFT assumptions disclaimed in footnote 3 and Section 4.3; the deferred second-level SPT obstruction could change the physical classification.","rationale":"The reader identified the same load-bearing weakness: the physical conclusions rely on the coarse SymTFT setup, explicitly disclaimed by the authors, and the second-level obstruction is deferred. I agree with that assessment. The theorem itself is a statement about Lagrangian subgroups and appears internally consistent; it is supported by detailed appendices and by reproducing prior results in the literature. The concern is not that the classification is wrong within its stated assumptions, but that the bridge from 'no invariant Lagrangian' to 'no trivially gapped symmetry-preserving phase' is exactly where the neglected SPT/spin/torsion data enter. Because the authors themselves flag this limitation and leave the refinement to future work, a conditional verdict is appropriate. My stress-test does not find an additional independent flaw in the Galois-theoretic argument, and I do not see grounds to reject or to upgrade to accept without the refinement. The verdict should remain CONDITIONAL as the reader set it.","tokens_in":65992,"tokens_out":10691,"duration_ms":119243,"concrete_test":"Take the S-duality example of Section 5.3.1 with N=5 (p=1 mod 4, k=1) and recompute the boundary conditions of F_A^G = Z_5^2 BF orbifolded by G=Z_4 using the methods of references [23,24], including the most general 5D SPT for the Z_4 action before gauging. Concretely, construct the |\\lambda;2> and |\\lambda^{-1};2> states and check whether the extra phase in the crossed action (4.37) can be absorbed by a 4D SPT for every choice of 5D SPT. If a trivially gapped phase disappears for some SPT choice, the physical corollary of Theorem 1 is incomplete as stated; if it survives all choices, the coarse no-go is supported in this representative case.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central physical claim is not Theorem 1 itself but the step that turns absence of a G-invariant Lagrangian subgroup of Z_N^{2r} into absence of a trivially gapped symmetry-preserving IR phase. That step presupposes that topological boundary conditions of the gauged SymTFT F_A^G are fully captured by Lagrangian subgroups plus the no-SPT orbifold datum used in Section 4.3. The authors are explicit that this is a coarse approximation: footnote 3 says they neglect spin structure, torsional homology, and work up to stacking of SPTs, and footnote 17 says that including a 5D SPT before gauging modifies the fusion algebra, potentially requiring a 4D SPT to absorb a phase, which is the 'second level obstruction' left to future work. Since that obstruction can forbid a symmetry-preserving gapped boundary even when an invariant Lagrangian exists, it can also change which claimed trivially gapped examples in Section 5 survive. Thus the existence criterion in Theorem 1 is necessary for the physical conclusion only within the stated coarse setup; the sufficiency direction for the actual gapped phases of the orbifolded theory is not established once SPT stacking is included. The theorem's internal Galois-theoretic content appears consistent and reproduces prior results, so the concern is with the physical interpretation, not with the algebra.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript develops a coarse classification of G-invariant Lagrangian subgroups of the finite abelian group Z_N^{2r}, which serve as topological boundary conditions of the 4+1d BF SymTFT for 3+1d theories with a 1-form symmetry and a non-invertible duality 0-form symmetry. The central result, Theorem 1 in §4.4, gives a cyclotomic-factorization criterion for whether a finite cyclic subgroup G⊂Sp(2r,Z) admits an irreducible G-invariant Lagrangian subgroup of Z_N^{2r}. The proof is carried out in Appendices B–D using p-adic methods, Hensel lifting, and Galois theory, with non-abelian extensions in Appendix E. The paper then argues that when no such invariant Lagrangian exists and the non-invertible symmetry is preserved along an RG flow, the IR cannot be trivially gapped: it must be either gapless or spontaneously break the symmetry. Applications are given to N=1∗ theories and to supersymmetry-breaking deformations of class S theories.","tokens_in":66204,"tokens_out":5261,"duration_ms":55571,"significance":"If the physical inference is accepted within its stated scope, Theorem 1 provides a genuinely useful algorithmic classification: it is simple to apply, reproduces earlier results in [20,22–24], and extends them to all finite cyclic subgroups of Sp(2r,Z) and to composite N. The appendices are detailed and include explicit worked examples for SL(2,Z), the Bolza surface, and the Klein quartic, which give strong consistency checks. The non-abelian extension in Appendix E is a valuable addition. The main caveat is that the gaplessness/SSB conclusions are conditional on the 'coarse' SymTFT setup: the authors explicitly neglect spin-structure/torsional-homology effects, SPT stacking, and the second-level obstruction mentioned in footnote 17. Within that scope the paper is a solid contribution; the internal algebra of Theorem 1 appears sound and is not circular.","major_comments":[{"comment":"The central physical step from 'no G-invariant Lagrangian in F_A' to 'no trivially gapped symmetry-preserving IR phase' is not fully established beyond the coarse setup. Footnote 17 explicitly defers the 'second level obstruction': including a 5D SPT before gauging modifies the fusion algebra and can require a 4D SPT to absorb a phase, which can forbid a gapped boundary even when an invariant Lagrangian exists. Thus Theorem 1 is a necessary condition within the SPT-free BF theory, but the sufficiency direction for actual gapped phases of the orbifolded SymTFT F_A^G is not proven. Because the title and abstract present protected gaplessness as a general conclusion, the paper needs either to prove the missing step, to restrict the physical claims explicitly to the SPT-free coarse setup, or to formulate the physical statement as a conditional theorem with the second-level obstruction as an explicit assumption.","section":"§4.3.2, footnote 17; §5.1–5.2"},{"comment":"The uplift argument from boundary conditions of F_A to boundary conditions of F_A^G assumes that every symmetry-preserving gapped boundary of the orbifolded theory is obtained either from a G-invariant boundary of F_A or from a collection of boundaries permuted by the D_g defects. This converse is not demonstrated; the Carqueville–Runkel–Schaumann orbifold datum is cited, but no proof is given that no 'intrinsically orbifold' symmetry-preserving boundary exists. If such a boundary existed, it would provide a trivially gapped phase even when no invariant Lagrangian exists, which would invalidate the claimed gaplessness conclusion. Please provide a proof of this lifting statement or explicitly state it as an assumption of the coarse classification.","section":"§5.1, Eqs. (5.3)–(5.7)"}],"minor_comments":[{"comment":"The symbols ·, ⊙, and ⊗ in Table 1 are explained only in the caption and are easy to misread; please add a legend in the main text, preferably with one worked example showing how a row is read.","section":"Table 1, §4.4"},{"comment":"The statement that the boundary condition B^{(2)}_{A+B} is 'not realized' by N=1∗ theories is imported from [46] and is not derived here; please mark it explicitly as an assumption or supply a direct dynamical argument.","section":"§5.3.1, N=2 example"},{"comment":"The sentence 'any one-form symmetry that is not Z_{p^k}^r is anomalous' is stronger than what is proved: the preceding argument shows that a torsional Lagrangian intersects every other Lagrangian, which obstructs a trivially gapped phase, but the identification of this as an 'anomaly' requires a more precise statement about the mixed anomaly. Please clarify the terminology.","section":"§5.2.1, after Eq. (5.16)"},{"comment":"The sentence 'Since its order modulo p is φ(m)' is confusing because the Euler totient φ(m) is not an order; please rephrase the argument about the order of ζ_n modulo p^k.","section":"Appendix G"},{"comment":"There are several minor language issues, e.g. 'As a first applications' in the abstract and 'a N−1 theory' in §5.3.2; a careful proofreading pass is recommended.","section":"Abstract and §1"}],"recommendation":"major_revision","confidential_remarks":"The paper is honest about its coarse assumptions, and the algebra behind Theorem 1 appears solid. The main risk is overclaiming in the title and abstract, given that the second-level SPT obstruction and the orbifold boundary-lifting step are not established. I do not see grounds for rejection, but the physical claims need to be either proven or carefully downgraded to conditional statements."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The punchline is that this paper earns its keep with Theorem 1, the cyclotomic obstruction criterion for G-invariant Lagrangian subgroups of Z_N^{2r} BF theories. That is new, it subsumes the special cases for S and ST_n in [22,23], and it visibly corrects at least one prior claim (the ST_2 row, where invariant Lagrangians exist for every N, not just perfect squares). The extension to the non-abelian examples GL(2,3) and GL(3,2) is a nice bonus and shows the method is not stuck on cyclic groups.\n\nWhat is good: the technical core is a real derivation, not a conjecture. The appendices build the p-adic and Galois machinery, Hensel lifting, and the reduction to Lagrangian subgroups carefully. The paper reports checks against prior literature and worked examples, and the table of n <= 35 is handy. The writing is honest about what is coarse: footnote 3 and Section 4.3 explicitly say they neglect spin structure, torsional homology, and SPT stacking, and footnote 17 names the second-level obstruction.\n\nThe soft spot is exactly the bridge from the algebra to the physics. Theorem 1 is a statement about Lagrangian subgroups of the BF lattice. The claim that 'no invariant Lagrangian => no trivially gapped symmetry-preserving IR phase' presupposes that gapped boundaries of the orbifolded SymTFT are fully classified by Lagrangian subgroups and by the no-SPT orbifold datum. The stress-test note is right that this is not established once SPT stacking is included; the second-level obstruction could forbid a symmetry-preserving gapped boundary even when an invariant Lagrangian exists, or allow one when the Lagrangian criterion says no. So the physical conclusions are conditional, not unconditional. The authors do not hide this, which is to their credit, but it means the headline 'protected gaplessness' should be read as 'protected gaplessness within the stated coarse SymTFT setup.'\n\nThe citation pattern looks fine. The self-citations are inputs to the framework, not fitted outputs. The proof is long and not machine-checked, so I would like a referee to poke at the Galois group statements in Appendix C and D, but on a careful read they are consistent.\n\nVerdict: send this to a serious referee. It deserves referee time. A specialist in SymTFT or generalized symmetries will get real value, and the table and Theorem 1 will be a standard reference for invariant Lagrangians even if some physical applications get refined later. I would not desk reject it; I would ask a referee to focus on the sufficiency direction and the second-level obstruction.","headline":"Theorem 1 is a genuine and useful classification result; the protected-gaplessness conclusion is conditional on coarse SymTFT assumptions the authors clearly flag.","tokens_in":66818,"tokens_out":2683,"would_cite":true,"duration_ms":27860,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves a number-theoretic criterion for when a duality rotation admits an invariant gapped boundary in the 4+1d BF SymTFT; absent such a boundary, a symmetry-preserving RG flow cannot end in a trivially gapped phase.","keywords":["non-invertible symmetries","duality defects","symmetry TFT","gaplessness","spontaneous symmetry breaking","Lagrangian subgroups","class S theories","topological orders"],"falsifier":"Look for a 3+1d theory whose non-invertible duality symmetry has a generator $g$ and level $N$ satisfying the obstruction of Theorem 1 (for example $S$-duality with $N=p^{2k+1}$, $p\\equiv 3\\pmod 4$), and compute the infrared Hilbert space of a symmetry-preserving deformation: finding a unique gapped symmetric vacuum, or a trivially gapped SPT that carries the symmetry, would falsify the no-go conclusion. A cheaper check is to enumerate Lagrangian subgroups of $\\mathbb{Z}_N^{2r}$ for a finite list of $(N,g)$ pairs and compare with Theorem 1's list; any mismatch disproves the criterion.","tokens_in":65760,"feed_emoji":"🔁","tokens_out":9330,"duration_ms":85093,"temperature":0.7,"pith_summary":"The paper asks which non-invertible duality symmetries of 3+1d quantum field theories can be realized by a trivially gapped, symmetry-preserving infrared phase. It works with the topological symmetry theory (SymTFT), a 4+1d BF gauge theory whose gapped boundaries are classified by Lagrangian subgroups of the charge lattice $A\\times A^\\vee$. The central result is Theorem 1: for a finite cyclic group $G\\subset Sp(2r,\\mathbb{Z})$ acting on the $\\mathbb{Z}_N^{2r}$ BF theory, a $G$-invariant Lagrangian subgroup exists unless an odd prime power $p_i^{k_i}$ and an odd cyclotomic power $\\Phi_{n_j}(x)^{m_j}$ in $\\det(x-g)$ satisfy $n_j$ coprime to $p_i$ with $-1\\in\\langle p_i\\rangle_{n_j}$ (with a special $p_i=2$ case). Whenever no invariant boundary exists, a symmetry-preserving RG flow cannot end in a trivially gapped phase: the IR is either gapless or the symmetry is spontaneously broken. The paper applies this to the $N=1^*$ theories and to class $S$ theories on higher-genus surfaces, and extends the criterion to non-abelian subgroups such as the automorphism groups of the Bolza surface and the Klein quartic.","feed_headline":"Duality defects rule out trivially gapped vacua in 4d","feed_subtitle":"A cyclotomic-polynomial test predicts which non-invertible symmetries leave the infrared gapless or break it.","key_machinery":"The load-bearing object is the topological symmetry theory itself: the 4+1d BF gauge theory $F_A$ with gauge group $A$, whose topological surface operators carry charges in $A\\times A^\\vee$ and whose gapped boundary conditions are Lagrangian subgroups $L\\subset A\\times A^\\vee$. A duality rotation $g\\in Sp(2r,\\mathbb{Z})$ acts on this charge lattice, and the existence of a symmetry-preserving gapped boundary is exactly the existence of a $G$-invariant Lagrangian subgroup. The technical engine that decides this is the cyclotomic factorization of $\\det(x-g)=\\prod_j\\Phi_{n_j}(x)^{m_j}$ together with Galois-theoretic and $p$-adic analysis (Hensel lifting and the structure of the Galois groups $\\langle p\\rangle_{n_j}\\subseteq \\mathbb{Z}_{n_j}^\\times$), which yield the simple obstruction $-1\\in\\langle p_i\\rangle_{n_j}$ of Theorem 1. For non-abelian $G$, the same question is settled by checking invariance under all generators after decomposing into irreducible subspaces over subfields of cyclotomic fields.","core_discovery":"On the paper's own terms, the discovery is a first coarse classification of the non-invertible zero-form duality symmetries that protect gaplessness or force spontaneous symmetry breaking in 3+1d. The classification is reduced to an arithmetic question about the finite group $G\\subseteq Sp(2r,\\mathbb{Z})$ acting on the $\\mathbb{Z}_N^{2r}$ BF symmetry theory: writing $\\det(x-g)=\\prod_j \\Phi_{n_j}(x)^{m_j}$ for the generator $g$, the theory has an irreducible topological $G$-invariant boundary condition precisely unless an odd power $p_i^{k_i}$ of $N$ and an odd multiplicity $m_j$ meet the obstruction of Theorem 1. If no such boundary condition exists and the non-invertible symmetry is preserved along the flow, the only symmetry-preserving IR options are gaplessness or spontaneous breaking; in particular, no SPT or trivially gapped topological order can carry the symmetry. The same SymTFT analysis also describes what happens when the symmetry is spontaneously broken, with the non-invertible defects acting as domain walls exchanging vacua.","pith_inferences":["Beyond the paper, the theorem's dependence only on the characteristic polynomial suggests the same $p$-adic Jordan-form method can be applied to infinite-order elements of $Sp(2r,\\mathbb{Z})$, producing constraints on duality symmetries that are not reductions of finite subgroups; the paper sketches this for $r=1$.","Beyond the paper, the neglected refinements named in footnote 3 and Section 4.3—spin structure, torsional homology, SPT stacking, and the second-level obstruction—could add invariant boundaries in some cases; redoing the classification with these data would tell whether any currently predicted 'gapless or broken' cases admit a trivially gapped escape.","Beyond the paper, the extra vacua corresponding to gauged duality symmetry (the $B^{(2)}$ boundary conditions) suggest that some symmetry-preserving flows may end on non-Lagrangian fixed points analogous to Argyres-Douglas theories; one could test this by looking for such vacua in non-SUSY deformations of Argyres-Douglas-like models.","Beyond the paper, the arithmetic condition can be read as a 4d analogue of a Lieb-Schultz-Mattis constraint: the non-invertible symmetry carries a charge that no topological line can realize, so a lattice or tensor-network construction preserving the duality defect should exhibit either gapless excitations or vacua exchanged by the defect."],"forward_implications":["For any finite abelian subgroup of $Sp(2r,\\mathbb{Z})$, one can algorithmically decide whether a 3+1d theory with that family of duality defects admits a trivially gapped symmetry-preserving IR phase by factoring the characteristic polynomial of the generator and checking the condition of Theorem 1.","If Theorem 1 rules out invariant lagrangians, preserving the non-invertible symmetry along a relevant deformation forces gaplessness or spontaneous symmetry breaking, including cases where the one-form symmetry itself may also break.","The $r=1$ analysis reproduces and extends known results for $SL(2,\\mathbb{Z})$ duality and triality defects in gauge theories, specifying exactly which $N$ admit invariant lagrangians.","Applied to $N=1^*$ theories, the boundary-condition analysis recovers the previously found vacuum structure and identifies an extra topological boundary condition corresponding to gauging $S$-duality, which may select global forms without ordinary Lagrangian descriptions.","For class $S$ theories on higher-genus surfaces, deformations that preserve cyclic duality symmetries exist; whether the IR must be gapless or spontaneously broken is decided by the prime factors of $N$ through the same cyclotomic criterion."],"supporting_citations":[{"why":"Classifies 4+1d topological orders and identifies the SymTFT as the relevant BF-type theory up to Witt equivalence.","marker":"[93]"},{"why":"Proves that all higher-dimensional defects in 4+1d topological orders are condensates of surfaces, justifying the Lagrangian-subgroup description of gapped boundaries.","marker":"[92]"},{"why":"Constructs the SymTFT for non-invertible duality defects, including the twist defects and the orbifolded theory $F_G$ used in Section 5.","marker":"[91]"},{"why":"Establishes that non-invertible symmetries can obstruct gapped phases, the effect this paper systematizes and extends.","marker":"[75]"},{"why":"Gives the prior anomaly analysis of non-invertible symmetries in 3+1d whose results are reproduced as special cases of Theorem 1.","marker":"[23]"},{"why":"Determines when duality defects are group-theoretical; those cases correspond to existence of invariant lagrangians and are recovered by the new criterion.","marker":"[22]"},{"why":"Supplies the $Sp(2r,\\mathbb{Z})$ action on class S theories and the invariant-subspace method over $\\mathbb{F}_p$ that the paper lifts to $\\mathbb{Z}_{p^k}$.","marker":"[20]"},{"why":"Provides the $N=1^*$ vacuum analysis that Section 5.3.1 recasts and refines from the SymTFT viewpoint.","marker":"[46]"},{"why":"Underlies Hensel lifting and the $p$-adic Galois statements used to turn the cyclotomic factorization into the obstruction condition.","marker":"[134]"}],"fun_headline_variants":["Cyclotomic test predicts gaplessness from duality symmetries","Non-invertible symmetries: gapless or broken, no trivial gap","Duality defects decide: gapless or symmetry broken","Cyclotomic arithmetic forces gapless or broken phases","Gaplessness protected by non-invertible duality defects"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion that no trivially gapped phase exists whenever no invariant Lagrangian exists presumes that every possible gapped boundary of the SymTFT is captured by a Lagrangian subgroup of the $\\mathbb{Z}_N^{2r}$ charge lattice, with spin-structure, torsional-homology, SPT-stacking, and second-level obstruction effects neglected.","fun_headline_variants_meta":{"raw":{"variants":["Cyclotomic test predicts gaplessness from duality symmetries","Non-invertible symmetries: gapless or broken, no trivial gap","Duality defects decide: gapless or symmetry broken","Cyclotomic arithmetic forces gapless or broken phases","Gaplessness protected by non-invertible duality defects"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001038,"raw_usage":{"total_tokens":4377,"prompt_tokens":963,"completion_tokens":3414,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":579,"completion_tokens_details":{"reasoning_tokens":3328}},"tokens_in":579,"tokens_out":3414,"duration_ms":24657,"temperature":1.0,"reasoning_tokens":3328,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:14:33.514069+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Look for a 3+1d theory whose non-invertible duality symmetry has a generator $g$ and level $N$ satisfying the obstruction of Theorem 1 (for example $S$-duality with $N=p^{2k+1}$, $p\\equiv 3\\pmod 4$), and compute the infrared Hilbert space of a symmetry-preserving deformation: finding a unique gapped symmetric vacuum, or a trivially gapped SPT that carries the symmetry, would falsify the no-go conclusion. A cheaper check is to enumerate Lagrangian subgroups of $\\mathbb{Z}_N^{2r}$ for a finite list of $(N,g)$ pairs and compare with Theorem 1's list; any mismatch disproves the criterion.","supporting_citations":[],"review_version":1}