{"id":"276c288e-8234-4047-9e12-48b0dc9e9fea","arxiv_id":"2505.05392","paper_version":2,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"The number of invariant factors of critical groups of arithmetical structures on a tree is bounded by a leaf-and-matching statistic, all trees whose arithmetical structures give only cyclic critical groups are classified, and every finite abelian group is realized as such a critical group on some…","lead":"This paper develops a generalized chip-firing game for arithmetical structures on graphs and uses it to control how many cyclic factors the associated critical groups can have. It classifies the trees on which every arithmetical structure has a cyclic critical group, and proves that every finite abelian group occurs as the critical group of some arithmetical structure on a tree.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the only soft point is Lemma 4.3, but its induction can be read as producing a fixed support set, so Theorem 4.5 goes through.","rationale":"I read the full manuscript with the central claims in view: the invariant-factor upper bounds from Theorem 4.5, the cyclic classification of Corollary 4.6, and the prescribed-critical-group constructions of Section 6. The reader's verdict of ACCEPT is justified. The reviewer's weakest assumption, Lemma 4.3, is indeed the place where the paper is most terse. The potential gap I probed is the passage from 'every divisor is equivalent to one nonzero at at most B+1 vertices' to the fixed-set condition of Theorem 2.4. If the support set may depend on the divisor, the Smith-normal-form argument would not apply. However, the induction in Lemma 4.3 can be made to produce a fixed set: fix one tentacle in each starlike piece to be zeroed, apply the recursive concentration on T', and then clear the central vertex of S1 using only tentacles other than the distinguished leaf v. This avoids firing at v after the T' induction, so the T' representative is not disturbed. The operations are linear over Z, so every integer chip count can be cleared by repeated borrow/fire, and the size count is exactly sum(ell_i-2)+1. Thus the use of Proposition 2.3 in Theorem 4.5(a) is sound under a natural strengthening of the lemma's statement. I also rechecked the additivity theorem, the star constructions, and the Euclidean tentacle construction in Proposition 6.3; the arithmetic there checks out, including the cancellation in the critical-group formula and the coprime-step argument for the long tentacle. The remaining issues are typos and terse passages, such as the example using 364 instead of 324 and the slightly compressed proof of Theorem 6.4 for beta < iota(T). These do not affect the main results. Since the central claims appear to hold, the verdict should remain ACCEPT.","tokens_in":20735,"tokens_out":51482,"duration_ms":490782,"concrete_test":"Take the smallest tree with two adjacent degree-3 vertices (an H-tree, so with two degree-3 vertices and four leaves) and its starlike decomposition into two 3-stars. For several small integer d-vectors, including all genuine arithmetical structures on this tree, enumerate all divisors modulo the column lattice L and check that every class has a representative supported on the fixed set Z consisting of one chosen leaf of the first star, one chosen leaf of the second star, and the merged vertex. Equivalently, verify via Proposition 2.3 that the gcd of the |V|-|Z| minors of the corresponding submatrix of L is 1. A failure would invalidate the fixed-set reading of Lemma 4.3 and hence Theorem 4.5(a).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central upper bounds rest on Lemma 4.3 and its use in Theorem 4.5(a). The apparent danger is that Theorem 2.4 requires a fixed set Z, while Lemma 4.3 as stated only guarantees support on some set of size at most B+1, and the induction sketch's 'borrow out along all other tentacles' could involve firings at the merged vertex v, disturbing the T' representative. I checked the proof with a fixed reading: choose a fixed tentacle of each S_i to zero, and after applying the induction to T', zero the central vertex of S1 using only the non-v tentacles rather than firing at v. This gives a fixed Z of size B+1 independent of the divisor, so Proposition 2.3 applies. The recursive step is linear over Z and never revisits an already-zeroed tentacle vertex. I therefore find no load-bearing error; the statement of Lemma 4.3 would benefit from saying 'fixed set of size at most...' but this is an expositional gap, not a correctness failure.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a generalized chip-firing framework for arithmetical structures and uses it to bound the number of invariant factors of the associated critical groups, with a focus on trees. The main results are: (1) an upper bound on the number of invariant factors in terms of a starlike decomposition of a tree (Theorem 4.5(a)); (2) a matching lower construction showing every number of invariant factors up to that bound occurs (Theorem 4.5(b)); (3) a classification of trees for which every arithmetical structure has cyclic critical group (Corollary 4.6); (4) a reinterpretation of the invariant-factor bound via a new tree invariant, the splitting irregularity number, and the 2-matching number (Theorem 5.5); and (5) constructions of arithmetical structures on broom graphs and on subdivisions of arbitrary trees with prescribed critical groups, culminating in the statement that every finite abelian group is realized as the critical group of some arithmetical structure on a tree (Section 6, especially Proposition 6.3 and Corollary 6.5). The paper also proves a wedge-sum additivity theorem for critical groups under a coprimality hypothesis (Theorem 3.3).","tokens_in":20886,"tokens_out":25937,"duration_ms":265757,"significance":"If the proofs are fully repaired, these are substantial results. The wedge-sum additivity theorem cleanly extends Lorenzini's p-primary result, the invariant-factor bounds give a new structural restriction on critical groups of arithmetical structures on trees, and the classification of trees with only cyclic critical groups is a natural and satisfying corollary. The constructive results in Section 6 are particularly strong: they not only realize every finite abelian group on a tree but do so on subdivisions of any given tree with enough leaves, and the broom construction is explicit and elementary. The paper is generally clearly written, the Smith normal form arguments in Section 2 are sound, and the authors are careful to identify the external results on which they rely, such as [18, Theorem 2.1] for starlike critical groups. The main weakness is that the central upper-bound lemma is not stated with the fixed-support property needed for its application, and its proof is too terse at the key inductive step.","major_comments":[{"comment":"The inference from Lemma 4.3 to Theorem 4.5(a) is not justified as written. Lemma 4.3 concludes only that, for each individual divisor, there exists a support set of size at most sum_i(ell_i - 2) + 1; this set may depend on the divisor. Theorem 2.4, however, requires a fixed set Z such that every divisor is equivalent to a divisor supported on Z. The proof of Theorem 4.5(a) invokes Theorem 2.4 immediately after Lemma 4.3, so the bound on invariant factors does not follow from the stated lemma. The lemma should be strengthened to assert the existence of a fixed set Z of size at most sum_i(ell_i - 2) + 1, with the proof adjusted accordingly, or Theorem 4.5(a) needs a different argument.","section":"Lemma 4.3 and Theorem 4.5(a)"},{"comment":"The induction step in the proof of Lemma 4.3 is under-specified at the merged vertex v. The induction hypothesis applied to T' allows arbitrary firings on T', including firings at v; when these firings are repeated on T, they change the chip count at the central vertex of S1. The subsequent instruction to 'borrow out along all other tentacles of S1' does not explain how the vertices of T' (which include v) are protected from disturbance, nor why the final support is contained in a fixed set independent of the original divisor. This is connected to the previous comment, but it deserves separate attention because the repair requires a detailed induction invariant: for example, a fixed support set for T' that includes v, and a rule that S1's central vertex is cleared using only non-v tentacles.","section":"Proof of Lemma 4.3"}],"minor_comments":[{"comment":"The statement of Lemma 4.3 says 'let d in Z^{V(G)}', but the graph is T; this should be 'd in Z^{V(T)}'.","section":"Lemma 4.3"},{"comment":"The phrase 'T has does not have non-adjacent vertices of degree at least 3' is grammatically incorrect, and the sentence 'By Theorem 4.5(b), it suffices...' should reference both Theorem 4.5(a) and Theorem 4.5(b), since the 'if' direction uses part (a) and the 'only if' direction uses part (b).","section":"Corollary 4.6"},{"comment":"In Example 6.2, the computation 'r(v4) = -(108 + 18 + 1) mod 364 = 197' uses 364, but r(v0) = 324; the modulus should be 324.","section":"Example 6.2 and Proposition 6.3"},{"comment":"In the statement of Theorem 6.4, the phrase 'an arithmetical structure on T'' should read 'an arithmetical structure on tilde T'; the current wording introduces a symbol T' that is not defined in the statement.","section":"Theorem 6.4"},{"comment":"In the final paragraph of the proof of Theorem 6.4, the tree obtained by subdividing all edges is denoted with the same symbol T as the original tree, which is confusing; using a different symbol such as tilde T would clarify the argument.","section":"Proof of Theorem 6.4"}],"recommendation":"major_revision","confidential_remarks":"The central theorems appear defensible, and I found no evidence of a false statement. The main issue is that Lemma 4.3, as stated and proved, does not supply the fixed support set required by Theorem 2.4; this is a load-bearing gap in the upper-bound arguments, although it seems repairable by a more careful inductive formulation. The paper also leans on the authors' own [2, Theorem 1] in Theorem 4.5(b), but that result is published and is also subsumed by the cited [18, Theorem 2.1], so I do not see a novelty concern. The constructive and classification results are significant enough to merit publication after the proof of Lemma 4.3 is made precise."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know about this paper: it settles two open-sounding questions in arithmetical structures on trees, and does so with clean chip-firing arguments. The headline results are Corollary 4.6, classifying which trees have only cyclic critical groups, and Corollary 6.5, that any finite abelian group with at most ℓ(T) − 2 invariant factors appears on some subdivision of T. The special case that every finite abelian group is realized on a tree is the paper's most striking claim.\n\nGenuinely new: Theorem 3.3 extends Lorenzini's wedge additivity from 'p does not divide the merged r-value' to full coprimality, and the proof is elementary and convincing. Theorem 2.4 gives a clean general bound on invariant factors via divisor concentration. Section 5's splitting irregularity number is a useful invariant, and the connection to 2-matching numbers (Theorem 5.5) recovers and extends Corrales–Valencia. I checked the main proof steps by hand, especially the kernel argument in Theorem 3.3 and the Euclidean tentacle construction in Proposition 6.3; they work.\n\nThe soft spot, as the stress-test notes, is Lemma 4.3, the divisor-concentration lemma underpinning Theorem 4.5(a). The proof is an induction sketch, and the statement says 'a divisor that is nonzero at at most ... vertices' without specifying a fixed support set. Theorem 2.4 needs a fixed Z. However, the stress-test's reading—choose a fixed tentacle to zero on each S_i, and after induction clear the merged vertex without firing there—produces a fixed set. The paper would be easier to referee if the lemma stated 'a fixed set of at most ... vertices.' This is an expositional gap, not a correctness flaw.\n\nCitation pattern is fine: [2] is a published result with its own proof, and the authors also attribute the star case to Lorenzini. The paper is honest about what is new. No invented entities or free parameters.\n\nWho is this for? Anyone working on chip-firing, sandpile groups, or arithmetical structures. It's a solid contribution that deserves a serious referee. I'd send it out. I'd cite the cyclic-tree classification and the realization theorem.\n\nRecommendation: accept after minor revision, with the referee pushing for a clearer Lemma 4.3.","headline":"Solid paper that settles two open questions in arithmetical structures on trees; the only soft spot is a terse induction in Lemma 4.3, which is fixable.","tokens_in":21493,"tokens_out":2834,"would_cite":true,"duration_ms":27877,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05C05","05C25","05C50","05C70","20K01"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that every finite abelian group is the critical group of some arithmetical structure on a tree, and that the number of invariant factors of such a group is bounded by a starlike decomposition of the tree.","keywords":["arithmetical structures","critical groups","chip firing","trees","invariant factors","sandpile group","starlike decomposition","2-matching number"],"falsifier":"Compute the Smith normal form of $\\operatorname{diag}(d)-A(T)$ for every arithmetical structure on a small tree $T$ with two non-adjacent degree-$3$ vertices; finding any critical group with more than $\\ell(T)-2-\\iota(T)$ invariant factors, or any divisor class that cannot be represented on at most that many plus one vertices, would refute Theorem 4.5 and Lemma 4.3.","tokens_in":2093,"feed_emoji":"🌳","tokens_out":6581,"duration_ms":111847,"temperature":0.7,"pith_summary":"An arithmetical structure on a graph assigns positive integer labels $r(v)$ to vertices so that each $r(v)$ divides the weighted sum of labels at neighboring vertices; the torsion part of the cokernel of $\\operatorname{diag}(d)-A(G)$ is its critical group, also known as the sandpile group or Jacobian. This paper uses a generalized chip-firing game, where firing at a vertex subtracts a column of that matrix, to bound the number of invariant factors of the critical group for trees in terms of a starlike decomposition into simpler trees. It proves that every finite abelian group is realized as the critical group of some arithmetical structure on a tree, and that if a tree $T$ has $\\ell(T)$ leaves, any group with at most $\\ell(T)-2$ invariant factors appears on a subdivision of $T$. It also classifies the trees on which every arithmetical structure has cyclic critical group: exactly paths, three-leaf starlike trees, and trees with two adjacent vertices of degree $3$ (and no vertex of degree at least $4$). A wedge-sum theorem shows that critical groups add when two arithmetical structures are merged at vertices whose $r$-values are coprime.","feed_headline":"Every finite abelian group is realized as a tree's critical group","feed_subtitle":"Generalized chip firing bounds how many factors the group can have and which trees force cyclic groups.","key_machinery":"Four pieces carry the argument. generalized chip firing: firing at a vertex $v$ replaces a divisor $\\delta$ by $\\delta$ minus the column of $\\operatorname{diag}(d)-A(G)$ indexed by $v$; divisors modulo firing form the cokernel, and the degree weighted by $r$ is preserved, so degree-zero classes form the critical group. starlike decompositions: a tree is recursively split into starlike trees (one vertex of degree at least $3$) and paths, and Lemma 4.3 uses borrowing along tentacles to concentrate any divisor at $\\sum_i(\\ell_i-2)+1$ vertices. the wedge-sum merge of Proposition 3.1, with the coprime-$r$ condition, gives $K(G_1\\vee G_2)\\cong K(G_1)\\oplus K(G_2)$ (Theorem 3.3). Finally, the starlike-tree formula $K(S)\\oplus (\\mathbb{Z}/r_0\\mathbb{Z})^2\\cong \\bigoplus_{i=1}^{\\ell} \\mathbb{Z}/d_i^*\\mathbb{Z}$ (Theorem 6.1) is the constructive engine: it turns prescribed invariant factors into $r$-values on a broom graph, and tentacles can be extended without changing the critical group.","core_discovery":"The central discovery is a transfer between chip-firing geometry and linear algebra: if every divisor on a graph is equivalent, under this generalized firing, to one supported on a set $Z$ of vertices, then the critical group has at most $|Z|-1$ invariant factors (Theorem 2.4). For trees, the paper shows that a starlike decomposition $\\{S_i\\}$ lets every divisor be supported on at most $\\sum_i(\\ell_i-2)+1$ vertices, giving the bound that every arithmetical structure on $T$ has at most $\\sum_i(\\ell_i-2)=\\ell(T)-2-\\iota(T)=|E(T)|-\\nu_2(T)$ invariant factors (Theorems 4.5 and 5.5). On the construction side, the paper combines a formula for critical groups of starlike trees with a wedge-sum additivity theorem: Theorem 6.1 computes critical groups of starlike trees, Theorem 3.3 makes critical groups additive under merging vertices with coprime $r$-values, and the resulting broom-graph construction realizes every finite abelian group; Corollary 6.5 realizes any group with at most $\\ell(T)-2$ invariant factors on a subdivision of any tree $T$. The classification of trees with only cyclic critical groups (Corollary 4.6) is a direct corollary of the invariant-factor bound: the allowed trees are paths, starlike trees with three leaves, and trees with exactly two adjacent degree-$3$ vertices and no vertex of degree at least $4$.","pith_inferences":["Because the support bound in Lemma 4.3 holds for arbitrary integer labelings $d$, not only for arithmetical-structure matrices, the same invariant-factor upper bound should apply to any integer matrix $\\operatorname{diag}(d)-A(T)$; a computational search over small trees with non-arithmetical $d$ could test whether the bound is sharp outside the intended setting.","The paper's bound ties invariant factors to purely graph-theoretic parameters (leaves, splitting irregularity number, 2-matching number), suggesting that the structure of the critical group is carried by branch vertices and their adjacencies; one could attempt analogous bounds for graphs of higher cyclomatic number by decomposing along blocks.","The realization result yields a natural optimization problem not addressed in the paper: the minimum number of leaves a tree needs to realize a given finite abelian group $G$, which Corollary 6.5 places between the number of invariant factors of $G$ plus $2$ and the leaf count of a chosen host tree; determining this minimum exactly would be a next step."],"forward_implications":["Every finite abelian group $G$ appears as the critical group of an arithmetical structure on a broom graph, that is, a starlike tree with at most one tentacle of length greater than one (Proposition 6.3).","For any tree $T$ and any finite abelian group $G$ with at most $\\ell(T)-2$ invariant factors, some subdivision of $T$ carries an arithmetical structure with critical group $G$ (Corollary 6.5).","The number of invariant factors of the critical group of any arithmetical structure on $T$ is at most $\\ell(T)-2-\\iota(T)=|E(T)|-\\nu_2(T)$, and every integer from $0$ up to that bound is attained by some arithmetical structure on $T$ (Theorems 4.5 and 5.5).","All arithmetical structures on a tree have cyclic critical group exactly when the tree is a path, a starlike tree with three leaves, or a tree with two adjacent degree-$3$ vertices and nothing of degree at least $4$ (Corollary 4.6 and Remark 4.7).","Under wedge merging with coprime $r$-values at the merged vertices, critical groups add as a direct sum; this also gives arithmetical structures on $c$-cyclic graphs with any prescribed finite abelian critical group, for every positive integer $c$ (Remark 6.7)."],"supporting_citations":[{"why":"Defines arithmetical structures, the generalized Laplacian, and gives the order formula for critical groups of trees used throughout the paper.","marker":"[16]"},{"why":"Provides the p-part wedge-sum result under a coprime condition that Theorem 3.3 extends to a full direct-sum isomorphism.","marker":"[19]"},{"why":"Determines critical groups of arithmetical structures on star graphs; Theorem 4.5(b) uses these star structures to realize $\\mathbb{Z}/2\\mathbb{Z}$^t as critical groups.","marker":"[2]"},{"why":"Supplies the critical-group formula for starlike trees (Theorem 6.1) that drives the broom-graph construction of prescribed critical groups.","marker":"[18]"},{"why":"Relates critical ideals of trees to 2-matching numbers; Proposition 5.4 and Theorem 5.5 recover and extend this relationship.","marker":"[9]"},{"why":"Supplies the generalized chip-firing framework for directed graphs that the paper adapts to arithmetical structures.","marker":"[3]"}],"fun_headline_variants":["Every finite abelian group is a tree's critical group","Tree critical groups realize all finite abelian groups","From chip firing to every finite abelian group","Only these trees make cyclic critical groups","Generalized chip firing classifies cyclic trees"],"cache_read_input_tokens":23680,"weakest_assumption_plain":"The load-bearing premise is Lemma 4.3's claim that, using borrowing along tentacles, every divisor on a tree can be concentrated on at most $\\sum_i(\\ell_i-2)+1$ vertices for arbitrary integer labelings $d$; if the induction silently leaves one extra vertex with chips, the invariant-factor upper bounds and the classification of cyclic critical groups collapse.","fun_headline_variants_meta":{"raw":{"variants":["Every finite abelian group is a tree's critical group","Tree critical groups realize all finite abelian groups","From chip firing to every finite abelian group","Only these trees make cyclic critical groups","Generalized chip firing classifies cyclic trees"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000789,"raw_usage":{"total_tokens":3511,"prompt_tokens":1010,"completion_tokens":2501,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":626,"completion_tokens_details":{"reasoning_tokens":2431}},"tokens_in":626,"tokens_out":2501,"duration_ms":17071,"temperature":1.0,"reasoning_tokens":2431,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T23:12:15.611624+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the Smith normal form of $\\operatorname{diag}(d)-A(T)$ for every arithmetical structure on a small tree $T$ with two non-adjacent degree-$3$ vertices; finding any critical group with more than $\\ell(T)-2-\\iota(T)$ invariant factors, or any divisor class that cannot be represented on at most that many plus one vertices, would refute Theorem 4.5 and Lemma 4.3.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines arithmetical structures, the generalized Laplacian, and gives the order formula for critical groups of trees used throughout the paper."},{"cited_title":"Reine Angew","cited_arxiv_id":null,"evidence_quote":"Provides the p-part wedge-sum result under a coprime condition that Theorem 3.3 extends to a full direct-sum isomorphism."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Determines critical groups of arithmetical structures on star graphs; Theorem 4.5(b) uses these star structures to realize $\\mathbb{Z}/2\\mathbb{Z}$^t as critical groups."},{"cited_title":"Reine Angew","cited_arxiv_id":null,"evidence_quote":"Supplies the critical-group formula for starlike trees (Theorem 6.1) that drives the broom-graph construction of prescribed critical groups."},{"cited_title":"Critical ideals of trees","cited_arxiv_id":"1504.06239","evidence_quote":"Relates critical ideals of trees to 2-matching numbers; Proposition 5.4 and Theorem 5.5 recover and extend this relationship."}],"review_version":1}