{"id":"1348aa21-8a47-4c59-942f-432271a2ae04","arxiv_id":"1908.10327","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Every regular infinite tree set is represented by the edge tree set of a tree-like space, and the regular tame ones are exactly the edge tree sets of ordinary trees.","lead":"This paper studies 'tree sets', abstract objects that capture tree-shaped structures in combinatorics, and extends their representation theory from finite to infinite objects. The authors show exactly which infinite tree sets come from ordinary trees, and introduce 'tree-like spaces' to represent all of them.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 4.13 rests on an unproved identification of P(u,v) as the unique pseudo-arc; compactness of this edge-chain union is the load-bearing gap.","rationale":"The reader's weakest_assumption identifies exactly the same load-bearing point: Lemma 4.13 depends on the assertion that P(u,v) is the unique pseudo-arc, a claim imported from [1] and not proved in the paper. My reading confirms that this is the most serious soft spot. The theorem would be true if the promised direct verification of compactness, connectedness, and the separation property for P(u,v) can be supplied; the construction is plausible and Lemma 4.12 already gives connectedness and edge-deletion disconnectedness, which via Proposition 4.9 would yield unique pseudo-arcs if that proposition's own external dependencies on [1, Remark 4.4 and Lemma 4.16] are accepted. But the paper's own footnote explicitly declines to introduce the needed machinery, and the direct verification is not carried out. Since Theorem 4.15(2) is the central claim, a CONDITIONAL verdict remains appropriate; my concern does not move the reader's verdict. Separately, the abstract's statement that 'every infinite tree set' is representable omits the regular hypothesis of Theorem 4.15, which is a presentation defect but not the mathematical gap identified here. No ad hominem is intended; the issue is proof completeness, not correctness.","tokens_in":17001,"tokens_out":8561,"duration_ms":93605,"concrete_test":"Carry out the deferred direct verification for Lemma 4.13 without citing [1]. (1) Special case: let τ be the tree set of Example 2.3 (an ω-chain →s1 < →s2 < ... < →t), let u = Op(Ðt) and v = Op(→t), and write out P(u,v) = ⋃{*e | →e ∈ v\\u}. Using the subbasis S(→e,r) of §4.4, prove P(u,v) is closed in Tpτq (hence compact), connected, and that P(u,v)-e separates u from v for every edge e of P(u,v). (2) General case: for an arbitrary chain C = v\\u between two consistent orientations, express P(u,v) as an intersection of subbasic closed sets, or exhibit a chain for which the union fails to be closed. If (1) or (2) fails, Lemma 4.13's order-preserving argument collapses; if both succeed, the representation theorem has independent support and the citation to [1] is harmless.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 4.15(2) is the paper's central representation claim, and its proof is Lemma 4.13. There, for vertices u,v of Tpτ1q, the set P(u,v) = ⋃{*e | →e ∈ v\\u} is asserted to be 'the unique pseudo-arc' in T with end-vertices u and v. The only support is the footnote: 'This follows immediately if one uses the machinery established in [1], which we do not introduce here. Alternatively one can show the connectedness of P(u,v) by repeating the proof that Tpτ1q is connected, and verifying the other properties of a pseudo-arc directly.' This is load-bearing: the order on τpTpτ1qq is defined by inclusions P(y,v) ⊆ P(x,v) ⊆ P(x,w), so if P(u,v) is not actually a pseudo-arc (in particular, compact), the map φ in Lemma 4.13 is not known to be order-preserving and the isomorphism τ1 ≅ τ(Tpτ1q) is unproved. Compactness is not automatic: P(u,v) is an arbitrary union of closed edges along a chain, and a graph-like space need not have all such unions closed. The promised direct verification is not carried out, and uniqueness of the pseudo-arc is likewise imported from [1]. Thus Theorem 4.15(2) depends on an external premise that the paper itself identifies but does not supply.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies representations of infinite tree sets, which are nested separation systems with no trivial elements. It proves two classification theorems: a tree set is isomorphic to the edge tree set of a graph-theoretic tree iff it is regular and tame (no chain of order type ω+1), and a tree set is isomorphic to the edge tree set of a 'tree-like space' (a compact graph-like space satisfying tree-like properties) iff it is regular. The proofs construct a tree T(τ) from splitting orientations for the tame case, and a tree-like space T(τ) from all consistent orientations for the general case, and establish the corresponding round-trip isomorphisms. The paper also proves monotonicity results for inclusions and minors. The main new content is the topological representation via tree-like spaces.","tokens_in":17254,"tokens_out":21376,"duration_ms":182589,"significance":"If correct, the results give a complete and clean classification of representable tree sets and introduce a natural topological notion of tree that accommodates limit edges. The constructions are explicit and several auxiliary results (e.g., compactness of T(τ), connectedness, and edge-deletion properties) are proved in the paper. The paper is a valuable bridge between abstract separation systems and graph-like spaces. However, the central representation theorem for tree-like spaces depends on an unproved pseudo-arc identification in Lemma 4.13, and the abstract and introduction overstate the results by omitting the regularity condition.","major_comments":[{"comment":"The proof of Lemma 4.13 asserts that for vertices u,v of T(τ1) the set P(u,v) = ⋃{*e | →e ∈ v\\u} is the unique pseudo-arc with end-vertices u and v. This assertion is not proved; the footnote either refers to machinery from [1] or promises a direct verification that is not carried out. In particular, compactness of P(u,v) is not automatic for an arbitrary union of closed edges along a chain, and without compactness P(u,v) is not a pseudo-arc. The subsequent inclusions P(y,v) ⊆ P(x,v) ⊆ P(x,w), which are used to prove that φ is order-preserving, therefore lack justification. Since Lemma 4.13 is the core of Theorem 4.15(2), the main representation theorem is incomplete. Please provide the missing proof or state precisely which result of [1] applies and verify its hypotheses.","section":"§4.5, Lemma 4.13 (footnote **)"},{"comment":"The abstract states that 'every infinite tree set' can be represented, and Theorem 2 in the introduction states that 'Every tree set is isomorphic to the edge tree set of a suitable tree-like space.' Similarly, Theorem 1 states that every tree set without a chain of order type ω+1 is isomorphic to the edge tree set of a tree. However, Theorem 3.9(1) requires the tree set to be regular and tame, and Theorem 4.15(1) requires regularity. Since edge tree sets of trees and tree-like spaces are always regular, non-regular tree sets (e.g., a single unoriented separation with →s < ←s) are counterexamples to these overstatements. The statements should be corrected to 'regular tame' and 'regular', respectively.","section":"Abstract and §1, Theorems 1 and 2"},{"comment":"The sub-base of the topology on T(τ) is not well-defined as written: the sets S(→e,r) are defined only for →e ∈ O1, but the sub-base is then taken over all →e ∈ τ. Moreover, the sets E+(→e) and E−(→e) are defined relative to the chosen orientation O1, and the claim that the topology is independent of this choice is not proved. Because T(τ) is the central object in Theorem 4.15, the definition must be made precise (for example, by defining S(→e,r) for every →e via the member of {→e,←e} that lies in O1) and the invariance should be proved.","section":"§4.4, definition of the sub-base"}],"minor_comments":[{"comment":"The notation *e for the closed edge (presumably ι_e([0,1])) is used but never defined; please define it.","section":"§4.4"},{"comment":"Lemma 4.18 duplicates Lemma 4.3 and is stated without proof; it should be removed or replaced by a reference.","section":"§4.5, Lemma 4.18"},{"comment":"The sentence 'Redoing the proof of Lemma 3.6 ...' is a very compressed argument for the order-preserving property; a direct argument would improve readability.","section":"§3.3, proof of Lemma 3.7"},{"comment":"Theorem 3.10 is said to be a special case of Theorems 4.16 and 4.17, but the reduction from tree-like spaces to graph-theoretic trees is not immediate and should be explained.","section":"§3.3, Theorem 3.10"},{"comment":"The assertion about the unique pseudo-arc in the contraction having point set {[x] ∈ T1 | x ∈ P(v,w)} is stated without proof; since this is a key step, a proof or a reference should be provided.","section":"§4.5, proof of Theorem 4.17"}],"recommendation":"major_revision","confidential_remarks":"The paper is interesting and likely correct, but the two major issues—the unproved pseudo-arc identification in Lemma 4.13 and the overstated theorems—need to be fixed. The ambiguity in the sub-base definition should also be resolved. I would be willing to review a revised version."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this paper. First, it proves something genuinely new: regular tame tree sets are exactly the edge tree sets of graph-theoretic trees, and regular tree sets are exactly the edge tree sets of the new \"tree-like spaces\" with limit edges. Second, the second theorem has a load-bearing gap in the proof as written, and the abstract overclaims by dropping the word \"regular\". Neither problem is fatal, but both need fixing before I would trust the paper as more than a strong preprint.\n\nSection 3 is the best part. The proof that a tree set without an ω+1 chain is represented by an actual tree is clean and mostly self-contained. The construction of T(τ) from splitting orientations is natural, the flip-lemma arguments are honest, and the bijection proofs in Lemmas 3.7 and 3.8 are convincing. This alone is a solid contribution, extending Diestel's finite result in exactly the right direction.\n\nSection 4 is more ambitious and more fragile. The topological constructions are sophisticated: the sub-base topology, the Alexander sub-base argument for compactness, and the connectedness proof are real mathematics and, as far as I can tell, correct. But Lemma 4.13, which carries Theorem 4.15(2), asserts that the union P(u,v) of edges along the chain v\\u is the unique pseudo-arc between u and v. The footnote says this follows from machinery in [1] or by repeating the connectedness proof, but neither the precise citation nor the direct verification is supplied. The stress-test is right: P(u,v) is an arbitrary union of closed edges, and compactness is not automatic in a graph-like space. If the cited machinery from [1] really covers this, fine, but the paper should say exactly which theorem in [1] does the work. As it stands, an editor or referee cannot check this step without reconstructing it.\n\nSmaller issues: the abstract says \"every infinite tree set\" where Theorem 4.15 says \"every regular tree set\"; that is a simple but real overstatement. There are also a few \"easy to check\" and \"not hard to verify\" moments in Lemmas 4.14 and 4.17 that are probably fine for the intended audience but could be tightened.\n\nWho is this for? People working in abstract separation systems, tangles, and infinite graph theory. The categorical flavour is new, and if the pseudo-arc gap gets fixed, this will be a useful and citable paper. I would send it to peer review, not desk-reject it, but I would ask the authors to make Lemma 4.13 verifiable and to correct the abstract.","headline":"Fresh, likely-correct extension of tree-set representation to infinite tree sets, with a real but repairable gap in the topological half and an abstract that overstates the theorem.","tokens_in":17789,"tokens_out":6239,"would_cite":true,"duration_ms":74562,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05C05","05C63","06A06"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves every regular tree set is the edge tree set of a tree-like space, with the tame case exactly the ordinary infinite trees.","keywords":["tree sets","separation systems","infinite trees","tree-like spaces","graph-like spaces","pseudo-arcs","limit edges","tame tree sets"],"falsifier":"Take a regular tree set containing a chain of order type $\\omega+1$ with no trivial separations and compute its space $T(\\tau)$. The theorem predicts that for the two endpoint orientations $u,v$ the set $P(u,v)$ is a compact connected pseudo-arc whose every edge separates $u$ from $v$. If for some such $\\tau$ the set $P(u,v)$ is disconnected, non-compact, or two distinct pseudo-arcs with the same endpoints appear, then Theorem 4.15(2) is false.","tokens_in":16779,"feed_emoji":"🌳","tokens_out":10309,"duration_ms":91135,"temperature":0.7,"pith_summary":"Tree sets are nested families of abstract separations that model tree-like structures in combinatorics, from tree-decompositions to tangles. Finite tree sets are already known to be representable as the edge tree sets of finite graph-theoretic trees; this paper extends that representation theory to infinite tree sets. It proves that a regular tree set is isomorphic to the edge tree set of an ordinary (possibly infinite) tree exactly when it contains no chain of order type $\\omega+1$ (no infinite ascending sequence capped by a top element), and that every regular tree set, without exception, is isomorphic to the edge tree set of a suitably constructed 'tree-like space'—a compact topological analogue of a tree that can possess limit edges. The upshot is a complete geometric model: every regular infinite tree set, however wild, is the edge tree set of some concrete tree-like object.","feed_headline":"Every infinite tree set has a tree-like representation","feed_subtitle":"Ordinary infinite trees cover exactly the tame regular tree sets; all others need limit edges.","key_machinery":"The load-bearing object is an order-geometric correspondence between a regular tree set $\\tau$ and a space $T(\\tau)$ whose vertices are the consistent orientations of $\\tau$ and whose edges are the unoriented separations. Each oriented separation $\\vec e$ is maximal in a unique consistent orientation $O(\\vec e)$; the edge corresponding to $\\vec e$ joins $O(\\vec e)$ to $O(\\bar e)$, and the partial order of $\\tau$ is recovered by a subbase of open sets $S(\\vec e,r)$ that point toward the vertex where $\\vec e$ is maximal. In the tame case only the splitting orientations are kept as vertices and the space reduces to an ordinary tree. The decisive technical claim is that the set $P(u,v)=\\bigcup\\{\\bar e:\\vec e\\in v\\setminus u\\}$ is the unique pseudo-arc between $u$ and $v$ in $T(\\tau)$, and this uniqueness is what turns the order-preserving bijection into an isomorphism of tree sets.","core_discovery":"The paper's central discovery is a pair of representation theorems. First, for a regular tree set $\\tau$ that is tame—no chain of order type $\\omega+1$—the consistent orientations that split $\\tau$ can be made the vertices of a graph-theoretic tree $T(\\tau)$, with edges the unoriented separations of $\\tau$; this tree is isomorphic to the original when $\\tau$ came from a tree, and $\\tau$ is isomorphic to $\\tau(T(\\tau))$. Second, for an arbitrary regular tree set, the same idea works with all consistent orientations as vertices and a topology added, yielding a compact connected graph-like space $T(\\tau)$ (a topological space built from vertices and edges, with a Hausdorff topology that allows limit edges) that is tree-like, in that any two vertices are joined by a unique pseudo-arc (a compact connected sub-space whose every edge is needed to separate its endpoints). Every regular $\\tau$ is isomorphic to the edge tree set of $T(\\tau)$, and every tree-like space is isomorphic to $T(\\tau(T))$. This overcomes the $\\omega+1$ obstruction: tree-like spaces may have limit edges, so chains of any order type can be represented.","pith_inferences":["If this representation is treated as canonical, then consistent orientations that are not splitting become limit points of the tree-like space, giving a concrete geometric reading of abstract tangles as 'directions' in the space.","The topology on $T(\\tau)$ suggests a natural notion of convergence of separations; proving a compactness theorem for this convergence could yield a geometric proof of tangle–tree duality for regular tree sets.","A testable extension: for tree sets with dense chains, the cofinality of the chain should control whether the corresponding pseudo-arcs are metric or non-metric, so computing the topological dimension of $T(\\tau)$ may classify tree sets by the shape of their limit chains."],"forward_implications":["Every regular tree set, even one with chains of order type $\\omega+1$, is isomorphic to the edge tree set of a compact connected tree-like space, so order-theoretic questions about such tree sets can be translated into topological questions.","The tame regular tree sets are exactly the edge tree sets of ordinary infinite trees, with a chain of order type $\\omega+1$ as the only obstruction.","Inclusion of regular tree sets corresponds to taking minors of the representing tree-like spaces, and minors of tree-like spaces correspond to subsets of their edge tree sets.","The constructions are idempotent up to isomorphism: $T(\\tau(T))\\cong T$ and $\\tau(T(\\tau))\\cong\\tau$ for every tree-like space $T$, with the analogous statements holding for tame tree sets and ordinary trees."],"supporting_citations":[{"why":"Supplies the graph-like space and pseudo-arc machinery, including the compactness and uniqueness claims used to identify $P(u,v)$ as the unique pseudo-arc in Lemma 4.13.","marker":"[1]"},{"why":"Provides the Extension Lemma 2.2, which the paper uses to define consistent orientations as vertices and to obtain unique maximal-element orientations.","marker":"[2]"},{"why":"Establishes the finite representation of tree sets by edge tree sets of trees and the basic theory of tree sets that the infinite extension builds on.","marker":"[3]"},{"why":"Introduces graph-like spaces as topological limit objects; tree-like spaces are defined by specialising this notion.","marker":"[10]"}],"fun_headline_variants":["Limit edges represent all infinite tree sets","Tree-like spaces represent every infinite tree set","Tame tree sets are trees; wild ones need limit edges","Infinite tree sets without ω+1 chains are trees"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The representation theorem for arbitrary regular tree sets rests on the imported pseudo-arc statement that, for any two consistent orientations $u,v$ of $\\tau$, the set $P(u,v)$ built in Lemma 4.13 is the unique pseudo-arc between them in $T(\\tau)$; if that set can fail to be compact, connected, or unique, the isomorphism $\\tau\\cong\\tau(T(\\tau))$ may collapse.","fun_headline_variants_meta":{"raw":{"variants":["Limit edges represent all infinite tree sets","Tree-like spaces represent every infinite tree set","Tame tree sets are trees; wild ones need limit edges","Infinite tree sets without ω+1 chains are trees"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000718,"raw_usage":{"total_tokens":3191,"prompt_tokens":876,"completion_tokens":2315,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":492,"completion_tokens_details":{"reasoning_tokens":2254}},"tokens_in":492,"tokens_out":2315,"duration_ms":17810,"temperature":1.0,"reasoning_tokens":2254,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:46:30.408582+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a regular tree set containing a chain of order type $\\omega+1$ with no trivial separations and compute its space $T(\\tau)$. The theorem predicts that for the two endpoint orientations $u,v$ the set $P(u,v)$ is a compact connected pseudo-arc whose every edge separates $u$ from $v$. If for some such $\\tau$ the set $P(u,v)$ is disconnected, non-compact, or two distinct pseudo-arcs with the same endpoints appear, then Theorem 4.15(2) is false.","supporting_citations":[{"cited_title":"Bowler, J","cited_arxiv_id":null,"evidence_quote":"Supplies the graph-like space and pseudo-arc machinery, including the compactness and uniqueness claims used to identify $P(u,v)$ as the unique pseudo-arc in Lemma 4.13."},{"cited_title":"Diestel,Abstract separation systems, Order 35 (2018), no","cited_arxiv_id":null,"evidence_quote":"Provides the Extension Lemma 2.2, which the paper uses to define consistent orientations as vertices and to obtain unique maximal-element orientations."},{"cited_title":"1, 171–192, DOI10.1007/s11083-017-9425-4","cited_arxiv_id":null,"evidence_quote":"Establishes the finite representation of tree sets by edge tree sets of trees and the basic theory of tree sets that the infinite extension builds on."},{"cited_title":"Thomassen and A","cited_arxiv_id":null,"evidence_quote":"Introduces graph-like spaces as topological limit objects; tree-like spaces are defined by specialising this notion."}],"review_version":1}