{"id":"a60b73a5-3319-4f60-8476-b6a678a838cd","arxiv_id":"1908.05371","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The relative L-invariant is zero only for the 4-ball among rational homology balls, and relative trisections are unique up to interior stabilization, relative stabilization, and the new relative double twist.","lead":"This paper defines a new measurement, the relative L-invariant, for smooth 4-manifolds with boundary, based on how complex their trisection diagrams must be. It shows that for rational homology balls, this measurement is zero exactly for the standard 4-ball, and it introduces new moves that relate any two trisections of the same manifold.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The proof of Theorem 4.5 depends on Claim 4.7, whose order-of-edges argument is unjustified: the inference that β_i∩γ_1≠∅ forces the edge E to precede E_{β_1} does not follow from the cut-system disjointness condition, leaving the induction unsupported.","rationale":"The central claim of the paper is Theorem 4.5, that rL(X)=0 detects B^4 among rational homology balls. The proof is by induction on genus, and the only step that genuinely carries the weight is Claim 4.6/4.7, which asserts that a minimal valid path descends to a minimal valid path after destabilization. The reader's weakest_assumption identifies exactly this. I agree: the proof of Claim 4.7 contains an unjustified order-of-edges inference. The claim is not a minor technicality; without it, the induction cannot proceed, and the if-and-only-if statement in the abstract is unproven. Other listed issues (duplicate Case 3 in Theorem 4.2, Remark 3.15's asserted values, hand-waving in Theorem 3.20) are real but peripheral to the main theorem. The verdict should remain CONDITIONAL: the gap may be repairable, but the proof as written is incomplete. A concrete check would settle whether the missing implication holds or whether there is a counterexample.","tokens_in":29043,"tokens_out":15705,"duration_ms":133986,"concrete_test":"Re-derive Claim 4.7 by writing out the cut systems along δ: for each edge, list the curves present and verify the disjointness condition precisely. Determine whether β_i∩γ_1≠∅ implies γ_j∩γ_1≠∅ under the labeling used after the slides; if not, construct a concrete genus-2, p=0,b=1 relative trisection diagram (e.g., a twice-stabilized B^4 diagram) and a minimal valid path δ where E occurs after E_{β_1} and the descended edge has intersection >1. This would falsify Claim 4.6 and break the induction in Theorem 4.5.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Claim 4.7 is the load-bearing step in the induction of Theorem 4.5. After destabilizing along α_g∪γ_g, the proof must show that every edge of the minimal path δ except the β_1-replacing edge descends to a type 1 edge. For an edge E replacing β_i with γ_j (i,j>1), the proof assumes that if the descended intersection |β'_i∩γ'_j| > 1, then β_i∩γ_1≠∅ and γ_j∩β_1≠∅. The contradiction is derived from edge ordering: γ_j∩β_1≠∅ implies E_{β_1} before E, while β_i∩γ_1≠∅ supposedly implies E before E_{β_1}. The second implication is not justified. A type 1 edge adding γ_j only requires γ_j to be disjoint from all curves already in the cut system except the one it replaces. If E_{β_1} has occurred, the cut system contains γ_1; this forbids adding a γ_j with γ_j∩γ_1≠∅, but it does not forbid adding γ_j merely because the replaced curve β_i intersects γ_1. No argument shows β_i∩γ_1≠∅ implies γ_j∩γ_1≠∅. Thus the purported contradiction is not established, and the induction step reduces the genus only if Claim 4.7 is true by some other argument, which is not supplied.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a relative L-invariant rL(X) for smooth, orientable, compact 4-manifolds with connected boundary, defined by measuring the lengths of certain paths in the p-cut complex HT_p(Σ) associated to a relative trisection surface. Two related invariants, rL∂(X) and rL◦(X), are also defined. The main theorem states that if X is a rational homology ball and rL(X)=0, then X is diffeomorphic to B^4, giving a trisection-theoretic analogue of the Kirby–Thompson result for closed 4-manifolds. The paper also proves a uniqueness theorem for relative trisections up to interior stabilization, relative stabilization, and a new relative double twist move, and gives an explicit algorithm for performing Murasugi sums of relatively trisected 4-manifolds.","tokens_in":29334,"tokens_out":18320,"duration_ms":166963,"significance":"If the main results are correct, this is a substantive contribution to trisection theory: it provides the first relative analogue of the L-invariant and shows that it detects the standard 4-ball among rational homology balls. The construction is self-contained, and the proofs are generally detailed. The Murasugi sum gluing algorithm and the strengthened uniqueness theorem for relative trisections are also valuable, as they address previously open structural questions. The paper also includes explicit computations and bounds relating rL∂ to the arc complex, which will be useful for future applications.","major_comments":[{"comment":"The case analysis in the proof of Theorem 4.2 is incomplete as written. The proof labels two different cases as \"Case 3\", and the second of these reads \"a0 changes to arc c0 in Aγ (which is not also in Aγ)\", which is self-contradictory. The intended case appears to be the one in which the arc changes once into the terminal Aα cut system, and the reduction to Case 2 by reversing the path is asserted without a precise formulation. Since Corollary 4.4 relies on Theorem 4.2 to produce lower bounds on rL, this proof must be completed before Theorem 4.2 can be considered established.","section":"Section 4, proof of Theorem 4.2"},{"comment":"The order-of-edges argument in Claim 4.7 is valid but too terse. The proof should explicitly state that every vertex of HT_p is a disjoint cut system and that a type 1 edge adds a curve which is disjoint from all curves in the cut system other than the one it replaces. With that convention, β_i ∩ γ_1 ≠ ∅ forces the edge E replacing β_i to precede the edge E_{β_1}, and γ_j ∩ β_1 ≠ ∅ forces E_{β_1} to precede E, giving the desired contradiction. No actual gap remains, but the exposition should be clarified so that future readers do not misread the argument as relying on an unjustified implication.","section":"Section 4, Claim 4.7"}],"minor_comments":[{"comment":"The phrase \"a path δ in HT0(Σ) from vα to vγ to vγ\" should read \"from vα to vβ to vγ\", since the path is composed of the α-to-β and β-to-γ segments.","section":"Section 4, proof of Theorem 4.5"},{"comment":"The numerical claims rL(T') = 2 and rL(T'') = 4 are asserted without computation or reference to a proof, and the sentence contains an unmatched closing parenthesis. Please either add the missing justification or clearly mark these as examples awaiting later results.","section":"Section 3, Remark 3.15"},{"comment":"The sentence \"By combining Lemma 5.5 and Proposition 5\" should refer to Proposition 5.3, not \"Proposition 5\".","section":"Section 5, Corollary 5.6"},{"comment":"In addition to the duplicate Case 3 label, the first Case 3 ends with a conclusion about the monodromy fixing an essential arc, but the text does not fully justify why sliding over β curves, after the α and β curves are made standard, implies isotopy in the α-page. This is likely fixable by a short argument, but it should be spelled out.","section":"Section 4, proof of Theorem 4.2"}],"recommendation":"major_revision","confidential_remarks":"The main theorem (Theorem 4.5) appears sound after careful reading; the alleged gap in Claim 4.7 is not genuine once one recalls that type 1 edges in the cut complex preserve disjointness of cut systems. However, the proof of Theorem 4.2 is not in a publishable state as written, and since Corollary 4.4 depends on it, the paper needs a moderate revision. The authors should also clean up several typos and missing justifications. I do not see a reason to doubt the novelty or significance of the work."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a substantial paper that deserves a real referee, but the proof of the main theorem has a gap that needs attention. The authors define rL(X), the relative analog of Kirby–Thompson's L-invariant, and prove that for rational homology balls, rL(X)=0 iff X is B^4 (Theorem 4.5). They also introduce a relative double twist, a uniqueness theorem for relative trisections (2.17) strengthening Gay–Kirby, and an explicit algorithm for Murasugi sum gluing (3.20). These are genuinely new, and the cut-complex framework is natural and carefully set up.\n\nThe soft spot is in the induction for Theorem 4.5. Claim 4.7 is load-bearing: it says that after destabilizing along a standard pair, the minimal path descends to type 1 edges. The proof uses an order-of-edges argument. The line 'Since β_i intersects γ_1, E must occur before E_{β_1}' is not justified. If E_{β_1} has already occurred, the cut system contains γ_1; the type 1 condition forbids adding a γ_j that intersects γ_1, but β_i∩γ_1 does not imply γ_j∩γ_1. The second half of the contradiction (γ_j∩β_1 forces E_{β_1} before E) is fine, but the first half doesn't follow. The stress-test note is right about this. I don't see a quick repair, though the overall strategy is plausible and the gap may be fixable. The authors need to either prove that implication or find a different descent.\n\nSmaller issues: Theorem 4.2 has two contradictory 'Case 3' labels; Remark 3.15 asserts exact values of rL without derivation; and the Murasugi-sum proof has some hand-waving in the final identification. None of these are fatal by themselves. The citation pattern is clean—several self-citations are to published, independent work.\n\nThis is a paper for trisection and 4-manifold topologists. It should go to peer review, with the Claim 4.7 issue flagged prominently. If the authors can fix that step, it's a solid contribution.","headline":"A valuable new relative invariant with genuinely new moves and an algorithm, but the proof of the main detection theorem has an unjustified ordering argument in Claim 4.7.","tokens_in":29905,"tokens_out":5083,"would_cite":true,"duration_ms":44863,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["57M99","57R15","57M15"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper introduces the relative $\\mathcal{L}$-invariant $r\\mathcal{L}(X)$ for compact 4-manifolds with boundary and proves that, for rational homology balls, $r\\mathcal{L}(X)=0$ exactly when $X$ is diffeomorphic to the standard 4-ball…","keywords":["relative L-invariant","trisections","4-manifolds with boundary","rational homology ball","cut complex","Murasugi sum","relative double twist","open book decomposition"],"falsifier":"Take any known contractible 4-manifold that is not diffeomorphic to the 4-ball, compute $r\\mathcal{L}$ from one of its relative trisection diagrams, and check whether the normalized minimum valid-path length is zero; a zero value for such a manifold would disprove the theorem. At the proof level, exhibit a valid path in which the curve $\\beta_i$ must be replaced by $\\gamma_j$ before $\\beta_1$ is replaced by $\\gamma_1$, contrary to the ordering used in Claim 4.7.","tokens_in":28825,"feed_emoji":"0️⃣","tokens_out":11816,"duration_ms":99631,"temperature":0.7,"pith_summary":"The paper introduces a numerical invariant, the relative $\\mathcal{L}$-invariant $r\\mathcal{L}(X)$, computed from shortest paths in a cut complex attached to a trisection surface of a compact 4-manifold $X$ with boundary. The main theorem states that if $X$ is a rational homology ball, then $r\\mathcal{L}(X)=0$ if and only if $X$ is diffeomorphic to the standard 4-ball. This gives a trisection-theoretic way to recognize the 4-ball among rational homology balls, parallel to how the closed-manifold version of the invariant detects the 4-sphere among rational homology spheres. The paper also supplies two structural tools: an explicit algorithm for gluing relatively trisected 4-manifolds by any Murasugi sum, and a proof that any two relative trisections of a given 4-manifold are related by interior stabilization, relative stabilization, and a new 'relative double twist' move.","feed_headline":"Zero of new invariant equals the 4-ball","feed_subtitle":"For rational homology balls, the relative L-invariant vanishes exactly when the manifold is diffeomorphic to the standard 4-ball.","key_machinery":"The central object is the relative $\\mathcal{L}$-invariant $r\\mathcal{L}(X)$, defined by minimizing, over relative trisection diagrams, the quantity $|\\delta|-3(g+p+b-1)+(k_1+k_2+k_3)$, where $\\delta$ is a valid path in the $p$-cut complex $\\mathcal{H}\\mathcal{T}_p(\\Sigma)$ of the trisection surface $\\Sigma$. In this complex, vertices are cut systems consisting of $g-p$ closed curves and $2p+b-1$ arcs; type-0 edges represent generalized handleslides, type-0$\\partial$ edges represent slides of arcs with common endpoints, and type-1 edges represent replacing a curve by one that intersects it in a single point. A path is valid if it travels through the $\\alpha$-, $\\beta$-, and $\\gamma$-components of the complex in order, with the minimal possible number of type-1 edges between good pairs (pairs of cut systems connected by exactly the algebraic minimum of type-1 edges). The other load-bearing construction is the relative double twist, a diagram-level move that realizes a $\\partial U$ move on the boundary open book; together with relative stabilization it connects any two relative trisections of a fixed 4-manifold.","core_discovery":"The central discovery is that the relative $\\mathcal{L}$-invariant $r\\mathcal{L}(X)$ detects exactly the standard 4-ball among rational homology balls. Concretely, $r\\mathcal{L}(X)=0$ for a rational homology ball $X$ if and only if $X\\cong B^4$; the forward direction is Theorem 4.5, and the reverse direction follows from Remark 3.7, where the trivial $(0,0;0,1)$-trisection of $B^4$ has an empty valid path and hence $r\\mathcal{L}(B^4)=0$. The invariant is defined as the minimum, over all relative trisections of $X$, of the normalized length of a valid path in the $p$-cut complex of the trisection surface; a zero value forces the minimizing path to consist entirely of type-1 edges connecting good pairs. The proof then reduces the diagram by destabilizing along a parallel pair of curves, inductively stripping away genus until the diagram must decompose as a connected sum of the trivial $B^4$ diagram and genus-1 $S^4$ trisections, which describes $B^4$.","pith_inferences":["Because $r\\mathcal{L}(X)=0$ characterizes the 4-ball among rational homology balls, the invariant can serve as an obstruction: a rational homology ball that is known not to be $B^4$ would need to have $r\\mathcal{L}(X)>0$, so computations on explicit trisection diagrams could certify nontriviality.","The Murasugi-sum algorithm suggests testable subadditivity properties: one might ask whether $r\\mathcal{L}(X\\# Y)\\le r\\mathcal{L}(X)+r\\mathcal{L}(Y)$ or whether $r\\mathcal{L}$ respects Murasugi sums of the boundary open books; the paper does not address these inequalities.","The relative double twist gives a trisection-level way to change the boundary open book by a $\\partial U$ move, so one could use it to define distances between relative trisections with different boundary data, or to convert any relative trisection into one inducing a prescribed open book on the same boundary 3-manifold; this goes beyond the paper's uniqueness statement.","A natural quantitative question the paper leaves open is whether $r\\mathcal{L}(X)=r\\mathcal{L}_\\circ(X)+r\\mathcal{L}_\\partial(X)$ (Question 3.12); if true, the interior and boundary contributions to the invariant would each be separately computable and could give finer information about 4-manifolds with small invariant."],"forward_implications":["For rational homology balls, the relative $\\mathcal{L}$-invariant is a complete detector of the 4-ball: $r\\mathcal{L}(X)=0$ if and only if $X\\cong B^4$.","For a closed rational homology 4-sphere $\\widehat{X}$, the closed $\\mathcal{L}$-invariant vanishes if and only if $\\widehat{X}\\cong S^4$ (Corollary 4.8).","Small boundary complexity forces the boundary to be a connected sum of $S^1\\times S^2$'s: $r\\mathcal{L}_\\partial(T)\\le 1$ implies $\\partial X\\cong \\#_{2p+b-1}S^1\\times S^2$, and below a related threshold $\\partial X$ has an $S^1\\times S^2$ summand.","There are 4-manifolds with arbitrarily large relative $\\mathcal{L}$-invariant, including families whose boundary homology stays bounded (Corollaries 4.4 and 5.8).","Relative trisections of a fixed 4-manifold are unique up to interior stabilization, relative stabilization, and relative double twists, removing the previous need to fix the boundary open book or assume rational-homology-sphere boundary."],"supporting_citations":[{"why":"Supplies the original $\\mathcal{L}$-invariant for closed 4-manifolds, whose definition and sphere-detection theorem the relative version extends.","marker":"[KT18]"},{"why":"Introduces relative trisections and proves existence and uniqueness up to interior stabilization for trisections inducing equivalent open books; this is the foundational framework used throughout.","marker":"[GK16]"},{"why":"Proves that every relative trisection is described by a diagram and gives the monodromy algorithm connecting diagrams to the boundary open book; the cut-complex viewpoint here is modeled on this.","marker":"[CGPC18a]"},{"why":"Provides the $\\partial U$ and Hopf-stabilization moves for open books that the new relative double twist transfers to the trisection setting and that Theorem 2.17 relies on.","marker":"[PZ18]"},{"why":"Establishes gluing of relatively trisected 4-manifolds with compatible open books; the Murasugi-sum algorithm extends this gluing.","marker":"[Cas16]"},{"why":"Shows that boundary diffeomorphisms of a 4-dimensional 1-handlebody extend over it, which is what lets a trisection diagram determine the trisected 4-manifold up to diffeomorphism.","marker":"[LP72]"}],"fun_headline_variants":["Relative L-invariant zero picks out B^4","L-invariant zero iff manifold is standard 4-ball","Zero relative L-invariant detects the 4-ball","Relative L-invariant: zero means diffeo to B^4","New invariant: zero only on the 4-ball"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that, after simplifying a minimal relative trisection diagram, the shortest path used to define the invariant survives the simplification without getting longer, and this survival depends on the order in which curves are exchanged along the path.","fun_headline_variants_meta":{"raw":{"variants":["Relative L-invariant zero picks out B^4","L-invariant zero iff manifold is standard 4-ball","Zero relative L-invariant detects the 4-ball","Relative L-invariant: zero means diffeo to B^4","New invariant: zero only on the 4-ball"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000407,"raw_usage":{"total_tokens":2156,"prompt_tokens":1026,"completion_tokens":1130,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":642,"completion_tokens_details":{"reasoning_tokens":1049}},"tokens_in":642,"tokens_out":1130,"duration_ms":9529,"temperature":1.0,"reasoning_tokens":1049,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:16:51.858894+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take any known contractible 4-manifold that is not diffeomorphic to the 4-ball, compute $r\\mathcal{L}$ from one of its relative trisection diagrams, and check whether the normalized minimum valid-path length is zero; a zero value for such a manifold would disprove the theorem. At the proof level, exhibit a valid path in which the curve $\\beta_i$ must be replaced by $\\gamma_j$ before $\\beta_1$ is replaced by $\\gamma_1$, contrary to the ordering used in Claim 4.7.","supporting_citations":[],"review_version":1}