{"id":"c19d13eb-8547-4b1f-b551-6dbc9a73c688","arxiv_id":"2607.14374","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"In globally hyperbolic spacetimes, the focalization time and the cut time are locally Lipschitz on their finite domains, yielding a Hausdorff-codimension bound for the cut locus.","lead":"This paper proves that, in globally hyperbolic spacetimes, the time when a causal geodesic stops maximizing (the cut time) is locally Lipschitz in timelike directions, with a controlled blow-up near the null cone. It also shows the cut locus has Hausdorff dimension at most n−1, extending classical Riemannian results.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Non-rigorous Lemma 4.2(iii) is the load-bearing gap: the quadratic Taylor estimate underpinning the main Lipschitz bound is asserted, not proved.","rationale":"The reader's verdict identified the null cut-point dichotomy (Theorem A.10) as the weakest assumption. That is a real gap, but it mainly affects null directions and continuity at null v*, not the main timelike result. The more fundamental problem is Lemma 4.2(iii), which the manuscript itself flags as 'not completely rigorous' and which is essential to the proof of Lemma 4.1 for every timelike pair. I therefore regard it as the single most load-bearing concern. The paper's central claim may still be true, and the gap appears fixable, so the verdict remains CONDITIONAL rather than REJECT. A full rigorous proof of Lemma 4.2(iii) would settle the matter.","tokens_in":24276,"tokens_out":8975,"duration_ms":82189,"concrete_test":"Write a complete, rigorous proof of Lemma 4.2(iii) for the compact set K = {(v,w): v∈U, ρ(v)=ρ(w), c_v(ρ(v))=c_w(ρ(w))} from Lemma A.17. In normal coordinates around x and around y, establish a uniform second-order Taylor expansion of φ∘exp_x(tv) with remainder r satisfying |r|+|∂_t r| ≤ C_K |v−w|^2, where C_K is independent of the test function φ as φ ranges over a bounded set of C^2 functions with controlled first and second derivatives. Then verify that the resulting bound on |X(ρ(v))| and |˙c_w(ρ(w))−˙c_v(ρ(v))−˙X(ρ(v))| holds with C_K independent of v,w in K. If the proof requires a factor depending on |v|_g^{-1}, the main theorem's estimate near the null cone fails.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The main theorem (Thm 1.4) rests on Lemma 4.1, whose proof uses Lemma 4.2(iii) at two critical points: to replace e1(w) by e1 with an error O(|v−w|^2), and to control ⟨Y, ˙X(ρ)⟩ after swapping covariant derivatives. Without the quadratic bounds (4.4)–(4.5), the factor sinh^2(ϕ/2) in (4.15)–(4.16) cannot be absorbed, and only a weaker (or divergent) Lipschitz estimate would follow.\n\nThe proof of Lemma 4.2(iii) is explicitly non-rigorous: it says 'we shall keep the argument intuitive, while being not completely rigorous' and then uses an arbitrary test function φ with a Taylor remainder whose constant 'depends only on φ'. This is not a minor omission: the passage from the scalar identity (4.7) to the vector bounds (4.4)–(4.5) requires uniformity of the remainder with respect to φ chosen from a bounded set of C^2 functions, and also uniformity over the compact set of (v,w) from Lemma A.17. The paper does not provide these uniformity arguments, nor does it address how the constant depends on |v|_g when v approaches the null cone—precisely where the denominator in (4.1) blows up.\n\nA further, secondary gap is the import of Theorem A.10 (null conjugate points imply non-maximality) from lecture notes without proof; this affects continuity of ρ at null v* and Corollary 1.6, but not the timelike core. Lemma 4.2(iii) is more load-bearing because it affects every timelike direction in Theorem 1.4.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies the cut time ρ and focalization time λ for a fixed point x in a globally hyperbolic spacetime. It claims: λ is locally Lipschitz on its relatively open domain (Thm 1.3); ρ is locally Lipschitz near timelike vectors with the quantitative estimate |ρ(v)−ρ(w)| ≤ C |v−w|_x / min{|v|^2_g, |w|^2_g} (Thm 1.4), with a refined result near null directions when ρ<λ (Cor 1.6); and dim_H Cut(x) ≤ n−1 (Cor 1.5). The method adapts Itoh–Tanaka's Riemannian argument, using index-form estimates, Jacobi fields, and Lorentzian subgradients. Appendices collect background on conjugate/cut points.","tokens_in":24664,"tokens_out":13150,"duration_ms":115230,"significance":"If complete, this is the first Lipschitz regularity result for the Lorentzian cut time and a natural extension of Itoh–Tanaka and Li–Nirenberg. The quantitative blow-up of the Lipschitz constant near the null cone is new and interesting. The paper is clearly written and mostly self-contained, with explicit constants and a systematic handling of compactness (Lemmas A.14, A.17). However, the central estimate Lemma 4.2(iii) is explicitly non-rigorous, and the null part depends on an unproved theorem from lecture notes. Thus the results are plausible but not yet established.","major_comments":[{"comment":"This lemma is load-bearing: Lemma 4.1 uses it at (4.15)–(4.16) to replace e1(w) by e1 − ˙X/|v| and to control ⟨Y, ˙X(ρ)⟩ after swapping covariant derivatives. The proof is explicitly non-rigorous ('we shall keep the argument intuitive, while being not completely rigorous'). The Taylor expansion (4.7)–(4.8) with an arbitrary test function φ does not establish the uniform vector-valued estimates (4.4)–(4.5): one needs a uniform remainder for a bounded family of C² functions and uniform control over the compact set from Lemma A.17, including the dependence on |v|_g as v approaches the null cone. Without this, the factor sinh²(φ/2) in (4.15)–(4.16) cannot be absorbed and Theorem 1.4 is not proved. A direct Taylor expansion of F(t,v)=exp_x(tv) should be written out.","section":"§4.1, Lemma 4.2(iii), Eqs. (4.4)–(4.5)"},{"comment":"Theorem A.10 (a null geodesic containing a conjugate point is not maximizing) is imported from lecture notes [2] without proof, and the text explicitly says the argument is omitted. This theorem is used in Theorem A.15, which in turn is used in Lemma 4.7 to produce the second distinct maximizing geodesic w(s0). It is also needed for Corollary 1.6 and the continuity of ρ at null vectors. Since the paper otherwise provides a self-contained survey, please either prove this theorem or replace [2] by a refereed published reference with a proof; currently the null-direction part of the main conclusions is unsupported.","section":"Appendix A, Theorem A.10 / proof of Theorem A.15"},{"comment":"The proof establishes (3.1) only for v,w∈U∩int(C_x), i.e. for pairs of timelike vectors. The theorem asserts local Lipschitz continuity on the whole relatively open domain, which includes null directions when v* is null. The phrase 'combining the continuity of λ with a connectedness argument' is not carried out. To justify the null case one needs an explicit limiting argument (e.g., approximate each causal vector by timelike vectors and pass to the limit using Lemma A.12), or a variant of Lemma 3.4 that does not require the denominator |v|_g to be nonzero. This is likely repairable, but it is not in the text.","section":"§3, proof of Theorem 1.3"},{"comment":"In the 'otherwise' case, C̃ is defined as C/ min{|v0|²_g,|v1|²_g} and then used as a Lipschitz constant for λ on U. As written, this constant depends on the pair (v0,v1) that was only fixed at the start of the argument; it is not a uniform constant on U. The proof should either fix a smaller neighbourhood with a uniform lower bound on |v|_g (when v* is timelike) or explicitly state that for each pair one applies the bound with the relevant minimum and then combines the inequalities. Please also check the consistency of the squared denominators in (4.1) and (4.2).","section":"§4, proof of Theorem 1.4 after (4.3)"}],"minor_comments":[{"comment":"Typos: 'Semiconvavity' in Proposition 1.2; 'Uniform estmates' in the heading of Lemma 4.2; 'costs no generality' in Appendix B should be 'costs no loss of generality'; in the proof of Theorem A.15 the curve is sometimes called γ and sometimes c.","section":"Throughout"},{"comment":"In the displayed Jacobi equation, the curvature term is written with E_j(t,v,w), although the parallel frame is E_j(t,v); please correct.","section":"Lemma A.4(i)"},{"comment":"The notation exp(y, −t ˙c_v(ρ(v))) is not defined; the map should be exp_y(tu). The derivative computation leading to the contradiction is very terse and should be written out, especially the role of the fact ρ(v_k)=ρ(w_k).","section":"Lemma 4.2(ii)"}],"recommendation":"major_revision","confidential_remarks":"I recommend major revision rather than rejection because the main strategy is sound and the missing pieces appear repairable. However, the admitted non-rigorous proof of Lemma 4.2(iii) and the unproved Theorem A.10 are serious; I would not be comfortable with acceptance until both are fixed. The self-citation [14] for semiconvexity of d is acceptable, but the editor may ask the author to give a fuller citation to [13] as well."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a genuine extension of Itoh–Tanaka, not a repackaging, and the timelike core is probably correct; but the written proof leans on a lemma that the author says is not completely rigorous, and the null-direction results depend on an unproved imported theorem. I'd send it to a referee, not desk-reject it, but I wouldn't cite the current version for the main theorem.\n\nWhat is actually new: first Lipschitz regularity for the cut time and focalization time in globally hyperbolic spacetimes, plus a clean Hausdorff-codimension corollary. The Lorentzian setting is not a translation: the non-openness of the cut-time domain, possible geodesic incompleteness, and the blow-up near the null cone all require new work. The author is upfront about which parts are adapted and which are new, and the exposition is generally careful.\n\nThe main soft spot is Lemma 4.2(iii). The statement is needed to control the quadratic error in the Jacobi-field estimate, and the proof says 'we shall keep the argument intuitive, while being not completely rigorous.' The Taylor expansion of phi composed with the exponential map gives a scalar remainder controlled by phi, but (4.4)–(4.5) require vector bounds uniform over a bounded set of test functions and over the compact set of pairs (v,w). That uniformity is not supplied. It is probably fixable—one can choose phi as a bounded linear functional with controlled Hessian—but as written it is a real gap, not a typo. It affects every timelike direction in Theorem 1.4, because the quadratic estimate is what absorbs the sinh^2(phi/2) factor.\n\nThe second gap is Theorem A.10, imported from lecture notes, which says a null geodesic containing a conjugate point is not maximizing. This is used for continuity of rho in null directions and for Corollary 1.6. The timelike core does not need it, but the advertised null-direction statements do. The self-citation to [14] is not a problem, since [13] is cited alongside it.\n\nFor a reader: this is aimed at Lorentzian geometers and anyone working on regularity of distance functions. I would bring it to a reading group, and I would want a careful referee on Lemma 4.2(iii). If the author closes that gap, I would cite it.","headline":"Credible extension of Itoh–Tanaka to Lorentzian geometry, but the main estimate has an explicitly non-rigorous gap that needs closing before I would rely on it.","tokens_in":25146,"tokens_out":6953,"would_cite":false,"duration_ms":68507,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53C50","58C07","49Q99"],"pacs":[],"model":"deepseek-v4-flash","headline":"The cut time is locally Lipschitz in globally hyperbolic spacetimes","keywords":["cut locus","cut time","focalization time","globally hyperbolic spacetime","Lipschitz continuity","Lorentzian geometry","Hausdorff dimension","conjugate points"],"falsifier":"Construct a globally hyperbolic spacetime with a future null geodesic from a point x that has a conjugate point before any cut point yet still maximizes Lorentzian length beyond the conjugate point. If such an example exists, Theorem A.10 is false, which would remove the cut-point dichotomy for null geodesics and invalidate the proof of the null-cone Lipschitz extension, while leaving the timelike theorem intact.","tokens_in":24148,"feed_emoji":"⏱️","tokens_out":5081,"duration_ms":48781,"temperature":0.7,"pith_summary":"This paper proves that two central functions in Lorentzian geometry, the cut time and the focalization time, are locally Lipschitz continuous, improving the previously known continuity results. For a fixed starting point in a globally hyperbolic spacetime, the focalization time is Lipschitz on its entire finite domain, including null directions. The cut time is Lipschitz in a neighborhood of every timelike direction whose geodesic is defined up to and including the cut time, with an explicit bound that blows up like the inverse square of the velocity near the null cone. Under an additional condition on the cut point, the Lipschitz property extends to null directions as well. A direct consequence is that the cut locus of a point has Hausdorff dimension at most n-1, matching the known Riemannian bound.","feed_headline":"Cut time is locally Lipschitz in globally hyperbolic spacetimes","feed_subtitle":"Extends the Riemannian regularity result; cut locus has Hausdorff dimension at most n-1.","key_machinery":"The central objects are the Lorentzian index form, evaluated on piecewise smooth vector fields orthogonal to a causal geodesic, and the corresponding Jacobi fields that encode how nearby geodesics deviate. Lemma 3.1 shows the index form is Lipschitz in the initial velocity, and Theorem 3.3 constructs vector fields with positive index near a conjugate time. For the cut time, the key geometric input is the cut-point dichotomy—that a cut point is either a first conjugate point or the endpoint of a second distinct maximizing geodesic. The proof also relies on the unit-hyperboloid estimate that connects the hyperbolic angle between two future unit timelike directions to the Euclidean distance bet","core_discovery":"The paper establishes the estimate |ρ(v)-ρ(w)| ≤ C|v-w|_x / min{|v|_g^2, |w|_g^2} for the cut time ρ, valid for future-directed timelike vectors v,w in a neighborhood of a given vector whose geodesic exists at the cut time. In particular, ρ is locally Lipschitz near such a timelike vector. The proof adapts a classical Riemannian strategy: it first shows the focalization time λ is locally Lipschitz everywhere by transferring control of the Lorentzian index form to nearby initial velocities; then it controls the derivative of ρ along line segments in the tangent space using the second maximizing geodesic delivered by the cut-point dichotomy. The denominator min{|v|^2,|w|^2} is inherent to the","pith_inferences":["The blow-up factor 1/min{|v|^2, |w|^2} is unlikely to be a technical artefact; it suggests that the cut time may be Hölder continuous with a specific exponent in directions approaching the null cone, a claim one could test explicitly in model spacetimes such as the Lorentzian cylinder.","If the imported theorem that a null geodesic containing a conjugate point is not maximizing turned out to fail, the argument for the null-cone part would collapse, while the timelike part would stand; this is a concrete open point worth checking.","The Hausdorff dimension bound opens the way to studying the rectifiability or smooth stratification of the cut locus in globally hyperbolic spacetimes, and to extending the argument to spacelike submanifolds, which the paper states as a natural next step.","A direct numerical check of the Lipschitz constant near the null cone in an explicit spacetime would sharpen the quantitative estimate and could reveal whether the constant C can be chosen independent of the base point in certain classes of spacetimes."],"forward_implications":["For any fixed point in a globally hyperbolic spacetime, the focalization time is locally Lipschitz on the open set where it is finite, including along null directions.","The cut time is locally Lipschitz in a neighborhood of every timelike direction whose geodesic is defined up to and including the cut time, with a uniform local Lipschitz constant.","If the cut point of a null direction occurs before its first conjugate point, then the cut time is locally Lipschitz in a full neighborhood of that null direction.","The cut locus of a point has Hausdorff dimension at most n-1, so it cannot fill an open set.","The cut time is locally semiconcave near timelike directions when the cut precedes the conjugate point, a property with independent consequences for the regularity of the time separation function."],"fun_headline_variants":["Cut time locally Lipschitz in globally hyperbolic spacetimes","Globally hyperbolic spacetimes: cut time is locally Lipschitz","Locally Lipschitz cut time in globally hyperbolic spacetimes","Cut time continuity upgraded to Lipschitz in globally hyperbolic spacetimes","New Lipschitz bound for cut time in globally hyperbolic spacetimes"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The proof relies on the theorem that a null geodesic containing a conjugate point is not maximizing, which is imported from lecture notes without proof; if that theorem is false, the cut-point dichotomy and hence the null-direction extension could fail.","fun_headline_variants_meta":{"raw":{"variants":["Cut time locally Lipschitz in globally hyperbolic spacetimes","Globally hyperbolic spacetimes: cut time is locally Lipschitz","Locally Lipschitz cut time in globally hyperbolic spacetimes","Cut time continuity upgraded to Lipschitz in globally hyperbolic spacetimes","New Lipschitz bound for cut time in globally hyperbolic spacetimes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.002746,"raw_usage":{"total_tokens":10287,"prompt_tokens":711,"completion_tokens":9576,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":455,"completion_tokens_details":{"reasoning_tokens":9482}},"tokens_in":455,"tokens_out":9576,"duration_ms":74352,"temperature":1.0,"reasoning_tokens":9482,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T02:16:15.265145+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct a globally hyperbolic spacetime with a future null geodesic from a point x that has a conjugate point before any cut point yet still maximizes Lorentzian length beyond the conjugate point. If such an example exists, Theorem A.10 is false, which would remove the cut-point dichotomy for null geodesics and invalidate the proof of the null-cone Lipschitz extension, while leaving the timelike theorem intact.","supporting_citations":[],"review_version":1}