{"id":"4bdbc066-0948-4c6a-83f0-dc05cce70c3b","arxiv_id":"2508.00767","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":7.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"Singular Soergel bimodules are recovered from Soergel bimodules through partial 2-categorical idempotent completions, and a local branching rule for gl_N foams yields a local version of Rose-Wedrich's decomposition theorem.","lead":"This paper applies higher idempotent completion to Soergel bimodules, showing that singular Soergel bimodules arise via partial 2-categorical idempotent completions. It also gives a local version of a known decomposition theorem for deformed coloured gl_N link homology.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Semistrict monoidal compatibility of the idempotent completion and the foam quotient is the load-bearing unproven premise; a minimal nontrivial type-A example would settle it.","rationale":"The reader's weakest_assumption already identifies the compatibility of semistrict monoidal structures with idempotent completion and the quotient functor as the pivotal premise. My stress-test agrees and sharpens it: the risk is not just the existence of semistrict structures, but the monoidal ideal property of the quotient. The recovery of singular Soergel bimodules by itself is a categorical construction that is often straightforward, but the identification with gl_N foams as semistrict monoidal 2-categories is a stronger claim requiring the quotient to respect the monoidal product. Because the full text is absent, no proof can be checked, so the verdict remains UNVERDICTED. I therefore recommend no change to the reader's verdict, but I flag this specific gap as the first thing to verify in a full review. The proposed concrete test is feasible: a complete computation for S_3 would either confirm the coherence conditions in the smallest nontrivial case or expose the failure. This is a good-faith concern, not a rejection; the paper may well contain the missing proof, but the abstract alone does not establish it.","tokens_in":646,"tokens_out":5278,"duration_ms":70326,"concrete_test":"Work out the smallest nontrivial type-A example, for instance the symmetric group S_3 with the two parabolic subsets {1} and {2}. Explicitly construct the partial idempotent completion of the 2-category of Soergel bimodules, define the monoidal tensor product on the newly added singular objects, and verify the associator and interchange coherence axioms of a semistrict monoidal 2-category. Then present the quotient functor to the corresponding gl_N foam 2-category and check that the ideal of 2-morphisms killed by the quotient is closed under tensor product with every object on both sides. If either the coherence axioms or the monoidal-ideal closure fails in this example, the paper's central identification cannot hold in general.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The abstract's central claim has two stages: (1) singular Soergel bimodules are recovered from Soergel bimodules via partial 2-categorical idempotent completion, and (2) in type A, the assembled 2-category is semistrict monoidal and 'certain quotients' are semistrict monoidal 2-categories of gl_N foams. Stage (1) is plausible but stage (2) carries the real weight: passing to an idempotent completion adds new objects that are formal splittings of idempotent 1-morphisms, and the monoidal product must be extended to these new objects. Even if the original 2-category is semistrict monoidal, the tensor product of two idempotents p and q only yields an idempotent up to a chosen 2-isomorphism; making the completion itself semistrict monoidal requires coherence conditions that are not automatic. More importantly, the quotient by the relations that produce the gl_N foam 2-category is monoidal only if the kernel is a monoidal ideal, i.e. closed under tensoring with all objects on both sides. The abstract does not state that this closure is proved. If either the semistrict structure on the completion or the monoidal ideal property of the quotient fails, the asserted identification with semistrict monoidal 2-categories of gl_N foams, and hence the fully local branching rule, would not follow. Since the full text is not available, this is not an accusation of error; it is a precise unresolved premise that the paper must establish.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The abstract announces two applications of higher idempotent completion: (1) a recovery of singular Soergel bimodules from Soergel bimodules via partial 2-categorical idempotent completions, and (2) in type A, an assembly of singular Soergel bimodules into a semistrict monoidal 2-category whose certain quotients are identified with semistrict monoidal 2-categories of gl_N foams, yielding a higher categorical branching rule and a fully local version of the Rose-Wedrich decomposition theorem for deformed coloured gl_N link homology.","tokens_in":970,"tokens_out":3376,"duration_ms":37806,"significance":"If established, the results would provide a unified higher-algebraic framework connecting Soergel bimodules and foam 2-categories, with potential implications for link homology and higher representation theory. The claimed fully local formulation of Rose-Wedrich is a substantive strengthening. The paper appears to build on established concepts rather than introducing ad-hoc axioms, and the abstract describes no free parameters. However, the absence of the full text makes it impossible to verify the key coherence and monoidal-ideal claims, so the significance assessment is provisional.","major_comments":[{"comment":"The claimed identification of 'certain quotients as semistrict monoidal 2-categories of gl_N foams' requires that the quotient functor be compatible with the semistrict monoidal structure. In particular, the tensor product on the idempotent completion of singular Soergel bimodules must extend to the new formal splitting objects with coherent 2-isomorphisms, and the kernel of the quotient functor must be a monoidal ideal (closed under tensoring with all objects on both sides). The abstract does not state that either property is proved; if either fails, the identification with gl_N foams and the local branching rule do not follow.","section":"Abstract (type-A paragraph)"},{"comment":"The recovery of singular Soergel bimodules from Soergel bimodules through 'partial 2-categorical idempotent completions' depends on the class of idempotents admitted by the completion and on the behavior of 2-morphisms under the completion. The abstract gives no indication of which class is used or how the completion is controlled, so the correctness of this reconstruction cannot be assessed from the manuscript as presented.","section":"Abstract (second sentence)"},{"comment":"The claim of a 'fully local version of Rose-Wedrich's decomposition theorem' is a strong assertion. To be convincing, the paper must exhibit a monoidal 2-functor or equivalence between the relevant completions/quotients and verify that the local branching rule assembles to the deformed coloured link homology decomposition. No such mechanism is visible in the abstract.","section":"Abstract (final sentence)"}],"minor_comments":[{"comment":"The phrase 'certain quotients' is vague; the authors should specify the generating 2-morphism relations of the quotient.","section":"Abstract"},{"comment":"The term 'partial 2-categorical idempotent completions' is not defined in the abstract; a precise definition or a reference to the authors' earlier work would help.","section":"Abstract"},{"comment":"The full text of the manuscript is not available for review; if this is an oversight, the complete version should be submitted.","section":"Submission format"}],"recommendation":"uncertain","confidential_remarks":"The manuscript appears to be an abstract-only submission. As a referee, I cannot verify the central claims without the full text. The specific load-bearing premise identified in my major comments—the semistrict monoidal compatibility of the idempotent completion and the monoidal ideal property of the quotient—is neither stated nor demonstrated in the abstract. I recommend requesting the complete manuscript before any further evaluation. If the full text was accidentally omitted, it should be supplied; otherwise, the submission is not reviewable in the normal sense."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know, first, that this is an abstract-only read, so any verdict is provisional. Second, the abstract makes concrete, checkable claims: singular Soergel bimodules are recovered from ordinary Soergel bimodules via partial 2-categorical idempotent completions, and in type A certain quotients of the assembled semistrict monoidal 2-category are identified with gl_N foam 2-categories. If those hold, the paper unifies two strands of the literature and gives a fully local Rose-Wedrich decomposition, which is a real advance.\n\nWhat the abstract does well: it states results plainly, names the relevant prior work (Rose-Wedrich, Lee-Gornik-Rasmussen-Wu), and doesn't oversell. The claimed applications to deformed coloured link homology are concrete enough that a reader can see what would be gained.\n\nThe soft spot is exactly where the stress-test points. Making an idempotent completion monoidal is not a formality: the new objects are formal splittings, and the tensor product of two idempotents is an idempotent only up to a chosen 2-isomorphism. The abstract doesn't state that the coherence is proved. Likewise, the quotient identifying the assembled 2-category with gl_N foams is monoidal only if the defining relations form a monoidal ideal, closed under tensoring on both sides. That closure is not mentioned. These are missing details, not detected errors. The abstract says 'certain quotients,' which is honest, but the referee will need to see the statements.\n\nA second limitation: without the full text, I can't judge the novelty against unpublished work or the soundness of the proofs. The results as stated don't appear in the cited literature, but that's a weak claim.\n\nIf I were handling this, I'd send it to a serious referee. The topic is active, the claims are substantive, and the construction has a clear checkable core: a minimal type-A example would resolve the monoidal-compatibility question quickly. That's the kind of paper that deserves referee time even if the final version needs heavy revision.","headline":"Promising abstract-only submission; the monoidal compatibility of the idempotent completion is the main thing to verify.","tokens_in":1416,"tokens_out":2416,"would_cite":false,"duration_ms":28276,"reading_group":"maybe","serious_thinker":"unclear","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["18N10","18M05","20F55"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper establishes that singular Soergel bimodules are recovered from Soergel bimodules via 2-categorical idempotent completions, which in type A assemble into semistrict monoidal 2-categories whose quotients are the gl_N foam…","keywords":["Soergel bimodules","singular Soergel bimodules","idempotent completion","2-categories","gl_N foams","link homology","branching rules","higher representation theory"],"falsifier":"To test the recovery claim, one could compute the spaces of morphisms of the idempotent completion between two explicit parabolic objects in a small type-A example (for instance $A_2$ or $A_3$) and compare them with the known singular Soergel bimodule morphism spaces; a single dimension mismatch would refute the claim. To test the foam identification, one could compare morphism spaces of the quotient 2-category with the known $\\mathfrak{gl}_N$ foam evaluations on an explicit pair of boundary webs, looking for an isomorphism mismatch.","tokens_in":499,"feed_emoji":"🔗","tokens_out":9856,"duration_ms":108913,"temperature":0.7,"pith_summary":"Soergel bimodules are algebraic objects, indexed by elements of a Coxeter group, that categorify the Hecke algebra and underlie several link homology theories; singular Soergel bimodules are the parabolic (relative) versions of these objects. This paper establishes that the singular versions are not new data: they are recovered from ordinary Soergel bimodules by partial 2-categorical idempotent completions. Specializing to type A, the paper assembles singular Soergel bimodules into a semistrict monoidal 2-category and identifies certain quotients of it with the semistrict monoidal 2-categories of $\\mathfrak{gl}_N$ foams, which are the surface-theoretic input for deformed coloured link homology. The paper then uses the same completion machinery to formulate a higher-categorical branching rule, giving a fully local version of Rose-Wedrich's decomposition theorem for deformed coloured $\\mathfrak{gl}_N$ link homology.","feed_headline":"Idempotent completion yields singular Soergel bimodules and gl_N foams","feed_subtitle":"This completion also yields a fully local branching rule for deformed colored gl_N link homology.","key_machinery":"The central device is the higher (2-categorical) idempotent completion: the generalization of the Karoubi envelope in which one formally adjoins splitting objects for idempotent morphisms, at the level of 1-morphisms or 2-morphisms of a 2-category. The paper applies this construction to the 2-category of Soergel bimodules, showing that the completed category carries the singular Soergel bimodules. In type A, the assembled semistrict monoidal 2-category (a 2-category whose tensor product is associative and unital up to coherent structure) is then quotiented and identified with the $\\mathfrak{gl}_N$ foam 2-category; the same completion operation is used to encode the branching rule that yields the local decomposition theorem.","core_discovery":"The central claim is that higher idempotent completion is the right bridge from ordinary to singular Soergel bimodules. Concretely, the author proves that partial 2-categorical idempotent completions of the 2-category of Soergel bimodules produce singular Soergel bimodules; in type A, the completed objects assemble into a semistrict monoidal 2-category, and the relevant quotients of that category are exactly the semistrict monoidal 2-categories of $\\mathfrak{gl}_N$ foams. A further application gives a higher-categorical branching rule for these foam theories, stated as a fully local version of Rose-Wedrich's decomposition theorem on deformed coloured link homology.","pith_inferences":["The same 'complete by idempotents first, then quotient' strategy is likely to generate singular versions of other categorical knot invariants, not only Soergel bimodules, whenever the needed idempotents can be identified.","The fully local branching rule suggests that deformed coloured link homology can be computed by evaluating filling surfaces on individual boundary webs and gluing the results, a potentially algorithmic route to computations in larger colour representations.","If the idempotent completion is compatible with tensor products in general, the result points toward a general principle: many 'singular' or 'parabolic' versions of representation-theoretic 2-categories are not extra input but formal idempotent splittings of the non-singular category."],"forward_implications":["Singular Soergel bimodules do not need to be constructed separately: they are formally built from ordinary Soergel bimodules by splitting idempotents, so constructions and invariants defined on Soergel bimodules transfer to the singular setting.","In type A, the foam 2-categories underlying $\\mathfrak{gl}_N$ link homology can be obtained as quotients of a monoidal 2-category assembled from singular Soergel bimodules.","The higher-categorical branching rule gives a fully local version of Rose-Wedrich's decomposition theorem, so the decomposition of deformed coloured link homology can be computed from local pieces rather than requiring a global diagram analysis.","The idempotent-completion formulation applies uniformly to the Lee-Gornik-Rasmussen-Wu deformations, packaging the deformation parameters into the same 2-categorical structure."],"supporting_citations":[],"fun_headline_variants":["Idempotent completion builds singular Soergel bimodules and gl_N foams","Higher idempotent completion yields gl_N foams and singular bimodules","Idempotent completion: from Soergel bimodules to gl_N foams","Local branching rule for gl_N foams from idempotent completion","Idempotent completion recovers singular Soergel bimodules and gl_N foams"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument rests on the assumption that the 2-category assembled from Soergel bimodules can be given a consistent way to combine objects and morphisms that survives the idempotent-completion step and the quotient step; without that coherence, the comparison with $\\mathfrak{gl}_N$ foams and the local branching rule would break down.","fun_headline_variants_meta":{"raw":{"variants":["Idempotent completion builds singular Soergel bimodules and gl_N foams","Higher idempotent completion yields gl_N foams and singular bimodules","Idempotent completion: from Soergel bimodules to gl_N foams","Local branching rule for gl_N foams from idempotent completion","Idempotent completion recovers singular Soergel bimodules and gl_N foams"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000652,"raw_usage":{"total_tokens":2938,"prompt_tokens":844,"completion_tokens":2094,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":460,"completion_tokens_details":{"reasoning_tokens":1983}},"tokens_in":460,"tokens_out":2094,"duration_ms":15860,"temperature":1.0,"reasoning_tokens":1983,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T05:56:04.822637+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"To test the recovery claim, one could compute the spaces of morphisms of the idempotent completion between two explicit parabolic objects in a small type-A example (for instance $A_2$ or $A_3$) and compare them with the known singular Soergel bimodule morphism spaces; a single dimension mismatch would refute the claim. To test the foam identification, one could compare morphism spaces of the quotient 2-category with the known $\\mathfrak{gl}_N$ foam evaluations on an explicit pair of boundary webs, looking for an isomorphism mismatch.","supporting_citations":[],"review_version":1}