{"id":"53654acc-53ff-43e0-ac5b-1276865c5e25","arxiv_id":"2412.10113","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"Every d-flag sortable simplicial complex is claimed to be vertex decomposable, and its associated toric and Rees algebras are Koszul, normal Cohen-Macaulay domains.","lead":"The paper shows that toric rings and Rees algebras built from d-flag sortable simplicial complexes are Koszul, normal Cohen-Macaulay domains, and that every such complex is claimed to be vertex decomposable. The results generalize known behavior of proper interval graphs to higher-dimensional unit-interval complexes and give algebraic characterizations for this class.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 4.1 rests on a false assertion: the deletion complex used in the induction is not interval for the unit-interval complex with maximal cliques [1,4] and [3,6] and r=2, so the vertex-decomposability proof collapses.","rationale":"The reader's weakest assumption is exactly the step I find load-bearing, and the supplied example is valid. The flaw is not a missing justification that a diligent reader could fill in: for the listed interval complex, the deletion construction produces a complex that violates the definition of interval, so the induction hypothesis in Theorem 4.1 cannot be invoked. The example is a unit-interval complex, hence the failure occurs inside the d-flag sortable setting of the abstract. Sections 1–3 are largely independent of Theorem 4.1, and their results may survive, but the vertex-decomposability and Cohen-Macaulay characterization is not established as written. This leaves the paper in the same CONDITIONAL position identified by the reader.","tokens_in":17963,"tokens_out":12873,"duration_ms":126925,"concrete_test":"Independently recompute the deletion step for the configuration B1 = [1,4], B2 = [3,6], r1 = r2 = 2, i = 4. Verify whether Φ_Del is interval by listing its facets and checking whether every 3-subset of the interval of each facet is again a facet. Since {3,4,5} is missing, the interval property fails, confirming that the proof's assertion is false. To separate proof failure from theorem failure, also run a vertex-decomposability check on Ind(Δ) for this example and nearby small interval complexes to see whether the statement of Theorem 4.1 might still be salvageable.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The inductive step of Theorem 4.1 hinges on the assertion, stated without proof, that the complexes Φ_Del = ⋃_{i∉B_j} Δ_j^{[r_j]} ∪ ⋃_{i∈B_j} (Del_{Δ_j}(i))^{[r_j]} and the analogous Φ_Lk are interval simplicial complexes; this is what makes the induction hypothesis applicable to Ind(Φ_Del) = DelΓ(i) and Ind(Φ_Lk) = LkΓ(i). The assertion is false. Take Δ with B1 = [1,4], B2 = [3,6], r1 = r2 = 2, so Δ is the pure 2-dimensional unit-interval complex whose facets are all 3-subsets of B1 and B2. In the proof, p = 1 and i = 4, and Φ_Del = {1,2,3}^{[2]} ∪ {3,5,6}^{[2]}, whose facets are {1,2,3} and {3,5,6}. For the facet {3,5,6}, the interval [3,6] is not a clique of Φ_Del: the 3-subset {3,4,5} is not a facet, because vertex 4 has been deleted from both components. Hence Φ_Del is not an interval simplicial complex. The induction cannot be applied as written. Since Theorem 4.1 is the sole support for Corollary 4.3, and hence for the abstract's vertex-decomposability claim and the Cohen-Macaulay characterization, this gap is load-bearing.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies d-flag sortable simplicial complexes Γ. It proves (Theorem 1.1) that such Γ are exactly independence complexes of unit-interval simplicial complexes Δ. Using this characterization and the theory of sorting orders, it shows that the Rees algebras of the facet ideals of pure skeletons are Koszul, normal Cohen-Macaulay domains (Corollary 1.6). For the toric ring R_Γ, the paper determines the a-invariant for the 2-flag (perfect graph) case (Proposition 2.1), recovers a Gorenstein characterization (Theorem 2.2), and for d>2 gives partial results on the divisor class group, canonical module, Gorenstein property, and a-invariant, some conditional on Conjecture 3.3. Finally, it claims that every interval simplicial complex has a vertex-decomposable independence complex (Theorem 4.1), yielding that every d-flag sortable complex is vertex decomposable and hence Cohen-Macaulay if and only if pure (Corollary 4.3).","tokens_in":18290,"tokens_out":18382,"duration_ms":167846,"significance":"If the main results hold, the paper provides a clean combinatorial characterization of Cohen-Macaulayness for d-flag sortable complexes and strong homological properties (Koszul, normal, Cohen-Macaulay) for the associated Rees and toric rings. The paper makes good use of existing machinery: it relies on established sorting orders, Gröbner basis theory, and the divisor class group descriptions from [13]. It also provides a new proof of a known Gorenstein characterization for toric rings of independence complexes of perfect graphs. The main caveat is that the vertex-decomposability theorem, which underpins the final characterization, rests on an unproved and nontrivial assertion in the proof of Theorem 4.1. Sections 1–3 are detailed and appear sound, though several statements for d>2 are conditional on Conjecture 3.3. Overall, the paper is valuable if the gap is fixed.","major_comments":[{"comment":"The proof asserts without proof that 'Del_Γ(i) and Lk_Γ(i) are interval simplicial complexes, as well', and this assertion is used to apply the induction hypothesis to the complexes inside Ind(·). This is load-bearing: without it, the induction step does not go through, and thus Theorem 4.1, Corollary 4.2, and Corollary 4.3 are not established. The issue is that after deleting vertex i, the vertex set is [n]\\{i}, which is not an initial segment, while the definition of interval simplicial complex in the paper is made for complexes with vertex set [n]. The authors should either prove the assertion after an explicit order-preserving relabeling of the vertex set and verification of the interval condition, or supply a different argument. The surrounding text suggests that the intended objects are the complexes Φ_Del and Φ_Lk, not the independence complexes themselves, so a clarification of the notation is also needed.","section":"Section 4, proof of Theorem 4.1"},{"comment":"The notation is confusing: Del_Γ(i) and Lk_Γ(i) are independence complexes, yet the sentence immediately after the display says they are interval simplicial complexes; presumably the complexes inside the Ind(·) operators are meant. Please rewrite for clarity and provide the missing proof that these inner complexes are interval simplicial complexes, or replace the argument with a different induction that does not require this assertion.","section":"Section 4, display before 'Notice that'"}],"minor_comments":[{"comment":"The term 'd-flag sortable simplicial complex' is used without a definition in the introduction; please define it explicitly, since it is central to the paper.","section":"Abstract / Introduction"},{"comment":"The indexing 'r_1, ..., r_n' appears to be a typo; there are m components, so the notation should be 'r_1, ..., r_m'.","section":"Section 4, equation (7)"},{"comment":"Reference [12] contains a typo in the title: 'tindependence ideals' should be 'independence ideals'.","section":"References"},{"comment":"The notation GΓ and Gc is easy to confuse; consider writing G^c for the complement graph.","section":"Proposition 3.7"}],"recommendation":"major_revision","confidential_remarks":"The paper is well-organized and demonstrates a strong command of the relevant literature. The main concern is the gap in the proof of Theorem 4.1. The counterexample in the review notes does not conclusively show the assertion is false, because after an order-preserving relabeling of the remaining vertex set, the example is an interval simplicial complex; nevertheless, the proof as written lacks the necessary justification, and the gap is load-bearing for the vertex-decomposability and Cohen-Macaulayness claims. I recommend requesting a rigorous proof of the deletion/link assertion, after which the paper will likely be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper’s real contribution is Theorem 1.1: a d-flag sortable complex is exactly the independence complex of a unit-interval simplicial complex. That characterization is clean and new, and it drives the toric Rees algebra results in Sections 1–3. The ℓ-exchange property, the quadratic Gröbner basis, and the divisor class group computations for d>2 are all substantial, and the authors are honest about Conjecture 3.3 limiting the full divisor class group determination. The radicality criterion for (t) in Corollary 3.8 is also a nice, self-contained result. If the paper stopped there, I would be enthusiastic.\n\nThe soft spot is Theorem 4.1, the claim that every interval simplicial complex has a vertex-decomposable independence complex. The proof says, without argument, that the deletion and link complexes are again interval simplicial complexes. That is not a one-line observation. After deleting a vertex, the complex sits on a subset of [n], not on [n] itself, and the induction hypothesis as stated applies to complexes on a full initial segment. You need an order-preserving relabeling, and you need the proper interval graph ordering to ensure that the intervals' right endpoints increase with the left endpoints. None of this appears in the proof. So the induction is not rigorous.\n\nThat said, the specific counterexample in the stress test with B1=[1,4], B2=[3,6], r=2 does not land. After deleting vertex 4, the vertex 4 is gone; the interval [3,6] inside the new complex is just {3,5,6}, which is a clique. The note treats the deleted vertex as still present, which is a misreading. So the theorem may well be true, but the paper does not prove it as written.\n\nWho is this for? People working on sortability, toric rings of independence complexes, and vertex decomposability. The core ideas are good enough that a serious referee should look at it, but the referee will likely send it back for a rewrite of Section 4. I would not cite the vertex-decomposability result until that is fixed. I would bring Sections 1–3 to a reading group, but not the current version of Section 4.\n\nRecommendation: send to peer review, with a clear request to scrutinize and repair Theorem 4.1.","headline":"Solid core in Sections 1–3, but Theorem 4.1 has a genuine technical gap that needs a rewrite; the stress-test counterexample, however, misreads the interval definition.","tokens_in":18856,"tokens_out":24046,"would_cite":false,"duration_ms":219612,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["13F55","13A30","13C14","05E40"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that every d-flag sortable simplicial complex is vertex decomposable, and that the associated toric and Rees algebras are Koszul, normal Cohen-Macaulay domains.","keywords":["sortable simplicial complexes","toric rings","Rees algebras","Cohen-Macaulay","vertex decomposable","divisor class group","unit-interval simplicial complexes","Gorenstein"],"falsifier":"Check the deletion step in Theorem 4.1 on a concrete interval complex: take the 2-dimensional unit-interval complex with maximal cliques $[1,4]$ and $[3,6]$, delete vertex $4$, and see whether the resulting deletion complex is still an interval simplicial complex; if it is not, the induction as written breaks down.","tokens_in":17715,"feed_emoji":"📐","tokens_out":10065,"duration_ms":93684,"temperature":0.7,"pith_summary":"Sortable simplicial complexes are rigid enough that their algebraic companions inherit strong properties, and the paper shows this rigidity pays off. The main claim is a characterization: a d-flag sortable complex is exactly the independence complex of a unit-interval simplicial complex, and every such complex is vertex decomposable, so it is Cohen-Macaulay if and only if it is pure. From this the authors conclude that the toric ring whose generators are the faces of the complex, and the Rees algebras of the facet ideals of its pure skeletons, are all Koszul, normal Cohen-Macaulay domains. Along the way they compute the divisor class group, canonical module, and a-invariant of these toric rings, and give combinatorial criteria for the Gorenstein property. If correct, the paper turns homological questions about these rings into purely combinatorial checks about intervals and cliques.","feed_headline":"Sortable simplicial complexes are vertex decomposable","feed_subtitle":"Rings built from these complexes are Koszul and normal; Cohen-Macaulayness reduces to purity.","key_machinery":"The load-bearing object is the unit-interval simplicial complex: a pure $(d-1)$-dimensional complex in which every facet $\\{i_1<\\cdots<i_d\\}$ has the whole interval $[i_1,i_d]$ as a clique, so its maximal cliques are intervals. The proof machinery is sortability: a set of face monomials is closed under the operation that orders the product of two monomials and alternately distributes the factors back, which yields quadratic reduced Gr\\\"obner bases for the toric rings and, via the $\\ell$-exchange property—a condition that a needed variable can be swapped into a sorted monomial—for the Rees algebras. For the divisor class group, the key identity is the support-form description of height-one monomial primes: each prime corresponds to a supporting hyperplane of the affine semigroup cone, and the Gorenstein property is read off from the coefficients of these forms. The final vertex-decomposability theorem is carried by an induction on interval complexes that deletes a well-chosen vertex and passes to the deletion and link.","core_discovery":"Let $\\Gamma$ be a $d$-flag sortable simplicial complex on $[n]$: its minimal non-faces all have size $d$, and the multiset of face monomials is closed under the sorting operation that orders the variables of a product and alternately distributes them back. The paper proves that $\\Gamma$ is sortable exactly when it is the independence complex $\\operatorname{Ind}(\\Delta)$ of a unit-interval simplicial complex $\\Delta$—a pure complex whose every facet $\\{i_1<\\cdots<i_d\\}$ spans an interval $[i_1,i_d]$ that is again a clique. Using this correspondence, the authors show that the toric ring $R_\\Gamma=K[x_F t : F\\in\\Gamma]$ and the Rees algebras of the facet ideals of the pure skeletons $I(\\Gamma^{[t]})$ are Koszul, normal Cohen-Macaulay domains; that all powers of $I(\\Gamma^{[t]})$ have linear resolutions and satisfy the strong persistence property; that the divisor class group of $R_\\Gamma$ is governed by height-one primes described by the maximal cliques of $\\Delta$; and that the $a$-invariant is bounded in terms of the clique number of $\\Delta$, with equality under a stated conjecture. The final theorem states that any interval simplicial complex $\\Delta$ has vertex-decomposable independence complex, so any $d$-flag sortable $\\Gamma$ is vertex decomposable and is Cohen-Macaulay exactly when it is pure.","pith_inferences":["Under Conjecture 3.3, the explicit height-one prime list would make the canonical module of $R_\\Gamma$ computable from the maximal cliques of $\\Delta$, turning Gorensteinness and the $a$-invariant into purely combinatorial checks.","The interval structure of $\\Delta$ suggests the Hilbert series of $R_\\Gamma$ might be expressible directly from interval lengths; computing examples where maximal cliques overlap would test both this and the conjecture.","If the vertex-decomposability theorem stands, it combines with standard shellability results to imply that the Stanley-Reisner rings of all interval simplicial complexes are shellable, extending the Cohen-Macaulay consequences beyond purity."],"forward_implications":["Every $d$-flag sortable simplicial complex $\\Gamma$ is vertex decomposable; in particular it is Cohen-Macaulay exactly when it is pure.","For such $\\Gamma$, both $R_\\Gamma$ and the Rees algebras $R(I(\\Gamma^{[t]}))$ are Koszul, normal Cohen-Macaulay domains, and all powers of $I(\\Gamma^{[t]})$ have linear resolutions and satisfy the strong persistence property.","When $\\Gamma$ is the independence complex of a perfect graph (the $d=2$ case), the $a$-invariant of $R_\\Gamma$ equals $-\\omega(G)-1$ and $R_\\Gamma$ is Gorenstein if and only if all maximal cliques of $G$ have the same size.","For $d>2$, Gorensteinness of $R_\\Gamma$ forces every maximal clique of $\\Delta$ to have size $2d-3$, and the $a$-invariant is bounded above by $-\\lceil\\omega(\\Delta)/(d-1)\\rceil$, with equality under Conjecture 3.3."],"supporting_citations":[{"why":"Sortability and sorting monomial orders, producing quadratic Gr\\\"obner bases for toric rings of sortable monomial sets.","marker":"[7]"},{"why":"Earlier paper on sortable simplicial complexes; established the flag case and the connection to proper interval graphs that this paper extends.","marker":"[12]"},{"why":"The $\\ell$-exchange property and the quadratic reduced Gr\\\"obner basis for Rees algebras of ideals with this property.","marker":"[10]"},{"why":"Provides the divisor class group, height-one monomial primes, and canonical module description for toric rings of simplicial complexes.","marker":"[13]"},{"why":"Support forms and the divisor-class-group criterion for the Gorenstein property used in Theorems 2.2 and 3.5.","marker":"[5]"},{"why":"Gorenstein characterization for toric rings of perfect graphs that Theorem 2.2 recovers with a new proof.","marker":"[19]"},{"why":"Vertex-decomposability of independence complexes of chordal graphs, the base case of the induction in Theorem 4.1.","marker":"[21]"},{"why":"Vertex decomposability and Alexander duality results used in Corollary 4.2 to derive the algebraic consequences.","marker":"[17]"}],"fun_headline_variants":["Sortable complexes: vertex decomposable, rings Koszul","Pure sortable complexes are exactly Cohen-Macaulay","Koszul and normal: toric rings of sortable complexes","Vertex decomposable: that's every sortable complex"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The final vertex-decomposability theorem rests on the claim that deleting a vertex from an interval simplicial complex, or passing to a link, again gives an interval simplicial complex; that preservation claim is what lets the induction in Theorem 4.1 keep going.","fun_headline_variants_meta":{"raw":{"variants":["Sortable complexes: vertex decomposable, rings Koszul","Pure sortable complexes are exactly Cohen-Macaulay","Koszul and normal: toric rings of sortable complexes","Vertex decomposable: that's every sortable complex"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000871,"raw_usage":{"total_tokens":3792,"prompt_tokens":988,"completion_tokens":2804,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":604,"completion_tokens_details":{"reasoning_tokens":2735}},"tokens_in":604,"tokens_out":2804,"duration_ms":24611,"temperature":1.0,"reasoning_tokens":2735,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T16:22:58.149070+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Check the deletion step in Theorem 4.1 on a concrete interval complex: take the 2-dimensional unit-interval complex with maximal cliques $[1,4]$ and $[3,6]$, delete vertex $4$, and see whether the resulting deletion complex is still an interval simplicial complex; if it is not, the induction as written breaks down.","supporting_citations":[{"cited_title":"Herzog, T","cited_arxiv_id":null,"evidence_quote":"Sortability and sorting monomial orders, producing quadratic Gr\\\"obner bases for toric rings of sortable monomial sets."},{"cited_title":"Herzog, F","cited_arxiv_id":null,"evidence_quote":"Earlier paper on sortable simplicial complexes; established the flag case and the connection to proper interval graphs that this paper extends."},{"cited_title":"Herzog, T","cited_arxiv_id":null,"evidence_quote":"The $\\ell$-exchange property and the quadratic reduced Gr\\\"obner basis for Rees algebras of ideals with this property."},{"cited_title":"Toric rings attached to simplicial complexes","cited_arxiv_id":"2302.03653","evidence_quote":"Provides the divisor class group, height-one monomial primes, and canonical module description for toric rings of simplicial complexes."},{"cited_title":"The toric ring of one dimensional simplicial complexes","cited_arxiv_id":"2306.05020","evidence_quote":"Support forms and the divisor-class-group criterion for the Gorenstein property used in Theorems 2.2 and 3.5."},{"cited_title":"Ohsugi, T","cited_arxiv_id":null,"evidence_quote":"Gorenstein characterization for toric rings of perfect graphs that Theorem 2.2 recovers with a new proof."},{"cited_title":"Woodroofe, Vertex decomposable graphs and obstruction s to shellability, Proc","cited_arxiv_id":null,"evidence_quote":"Vertex-decomposability of independence complexes of chordal graphs, the base case of the induction in Theorem 4.1."},{"cited_title":"Moradi, F","cited_arxiv_id":null,"evidence_quote":"Vertex decomposability and Alexander duality results used in Corollary 4.2 to derive the algebraic consequences."}],"review_version":1}