{"id":"9903f67d-3160-407e-bed1-0718f414d228","arxiv_id":"2411.13446","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Quasistatic fracture evolutions in nonlinear elasticity converge, after rescaling, to quasistatic crack growth in linear elasticity in two dimensions, without any a priori assumptions on the crack set.","lead":"This paper proves that quasistatic crack growth in nonlinear elastic materials converges, in the limit of infinite stiffness, to quasistatic crack growth in linear elasticity, with no restrictions on the crack geometry. It is the evolutionary counterpart of a known static linearization result and removes previous technical restrictions on the crack shape.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified; the nonsimple-material hypothesis is an explicit scope limitation, not a proof gap.","rationale":"The paper proves a precise evolutionary linearization theorem for the nonsimple model (2.4): any sequence of time-discrete minimizers has a subsequence converging, in the rescaled sense, to a linear quasistatic fracture evolution, with separate convergence of elastic and crack energies. The proof is long but each major step is either cited from [FS18] and [Fri20] or shown in detail. I focused on the rotation-decomposition and cut-off construction because they are the technical heart; the exponent estimates in Lemma 4.3 are consistent (β>2/3 makes εη_ε^3→0 and ε^{1−γ}η_ε→∞), and the refined jump-transfer lemma is a plausible adaptation of [FS18]. The minimality transfer and the energy-balance lower estimate handle broken-off pieces via Lemma 6.1, which is convincing. The density theorem in Appendix B is compressed, but its construction appears standard and the stated convergence is plausible; this is the step I would most want to see expanded, which is why I chose it as the verification check. The only true limitation is the nonsimple-material assumption, and this is an explicit hypothesis, not a hidden gap. The paper also clearly declares the restrictions to two dimensions, the lack of non-interpenetration, and the fact that it treats discrete-time approximate solutions rather than continuous-time nonlinear evolutions. Therefore the reader's ACCEPT verdict is appropriate, and I do not recommend any change.","tokens_in":45044,"tokens_out":30443,"duration_ms":305595,"concrete_test":"Expand the proof of the density theorem (Theorem 3.2) in Appendix B: verify that the partition-of-unity construction w_δ = Σ φ_i z_i^δ introduces no uncontrolled jump set on the overlapping collars beyond those counted in (B.10), and that the H^1(Jw_δ △ Jv) bound indeed tends to 0 as δ→0 for fixed η; if uncontrolled jumps appear, the reduction to regular competitors in Lemma 5.4 and Corollary 5.5 would be invalid.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The theorem's central claim—that time-discrete nonlinear fracture evolutions converge to a linear Griffith evolution—is conditional on the nonsimple-material assumption (2.4) with Hessian exponent β ∈ (2/3,1). This regularization is genuinely load-bearing: the rotation-decomposition in §4.2 (properties (R1)–(R5)) and the cut-off construction in Lemma 4.3 require the Hessian bound to control the set {|∇u_aux| ≥ η_ε}, and a purely first-gradient model would likely break these arguments. However, this is an explicit hypothesis of Theorem 2.2, the paper states the caveat in the introduction and §4, and Corollary 5.5 shows the Hessian contribution vanishes as ε→0. Hence the assumption narrows the theorem's scope to nonsimple (second-gradient) materials but does not invalidate the result as stated. I found no internal inconsistency or missing step in the proof chain; the remaining caveats (two-dimensionality, no non-interpenetration, and the use of discrete-time approximate solutions rather than true nonlinear evolutions) are all explicitly declared and do not affect correctness within the stated setting.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves an evolutionary linearization theorem for quasistatic fracture in two dimensions. For nonsimple nonlinear energies with a second-gradient penalization of order ε^{-2β}, β∈(2/3,1), the time-discrete approximate evolutions y_ε defined in (2.8) are shown to converge, after rescaling (y_ε−id)/ε, to a linear quasistatic Griffith evolution in the sense of Definition 2.1. The limit crack is exactly the union of the jump sets of the limit displacement at a countable dense set of times, and the total energies converge. No a priori assumptions on the geometry of the crack set are needed. The proof combines a priori estimates, a rotation decomposition, a cut-off construction, GSBD compactness and jump transfer, stability of unilateral minimality, and an approximate energy balance.","tokens_in":45233,"tokens_out":20595,"duration_ms":220098,"significance":"If correct, this is a substantial advance: it removes the restrictive crack-geometry assumptions of Negri–Zanini and Negri–Toader and supplies the evolutionary counterpart of the static Γ-limit in [Fri20]. The paper is unusually honest about its scope: the nonsimple-material hypothesis, the restriction to two dimensions, the absence of non-interpenetration, and the use of time-discrete approximate solutions rather than true nonlinear evolutions are all stated explicitly. The proof is organized as a detailed chain of lemmas with explicit rates and constants, and the cited external results are used as tools rather than as equivalent versions of the target claim. I found no circularity and no load-bearing gap.","major_comments":[],"minor_comments":[{"comment":"The displayed identity \"Jφ \\ Jˆu(t) = J_{ˆu(t)+φ} \\ Jˆu(t)\" is not true in general; only the inclusion \"⊆\" is needed and is valid, since a point where φ has a jump and ˆu has no jump is necessarily a jump point of ˆu+φ. Please replace the equality by the inclusion.","section":"§5.3, proof of Lemma 5.4"},{"comment":"The proof that the reflected extension satisfies H^1(J∇φ̂ \\ Jφ̂)=0 is compressed to a single sentence. Since this property is essential for the refined jump transfer lemma, please expand the argument: after reflecting, the jump set of ∇φ̂ is contained in the reflected copies of J∇φ, and the hypothesis H^1(J∇φ \\ Jφ)=0 ensures these are contained in the reflected copies of Jφ, which are contained in Jφ̂.","section":"§3, Lemma 3.4"},{"comment":"The statement says \"up to a subsequence in ε (not relabeled), y_ε ↝ u\", but the proof establishes the convergence for each t along possibly t-dependent subsequences and then upgrades it to the whole sequence using uniqueness of the limit. If this is intended, the wording could be made more precise so that readers do not expect a single subsequence working simultaneously for all t without the later identification argument.","section":"§2.4, Theorem 2.2"},{"comment":"The diagonal argument that produces the sequence n(ε)→∞ is only sketched. A short explicit description of how n(ε) is chosen after combining (5.24) and (5.25) would improve the readability of this key convergence step.","section":"§5.4, proof of Corollary 5.5"},{"comment":"The abstract says that crack growth in nonlinear elasticity is approximated by crack growth in linear elasticity, while the precise result concerns time-discrete approximate evolutions. The introduction later clarifies this point, but the abstract could be slightly more explicit that the convergence is proved for the discrete-in-time approximate evolutions defined in (2.8).","section":"§1 and Abstract"}],"recommendation":"minor_revision","confidential_remarks":"This is a strong paper with a careful proof and an explicit, honest list of limitations. The only issues I found are local presentation points, including a false equality that should be replaced by the needed inclusion. The central claim is sound within the stated hypotheses, and the paper is well suited to the journal's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The punchline: this is the right paper at the right time. It gives the evolutionary counterpart of Friedrich's static linearization result for brittle fracture, and it does so without any a priori crack-geometry assumptions, which is exactly what was missing since Negri-Zanini and Negri-Toader. If the main theorem is correct—and I saw no fatal gap—this closes a natural open problem in the linearization program.\n\nWhat is genuinely new: the convergence of time-discrete nonlinear quasistatic evolutions to a linear Griffith evolution, with simultaneous convergence of rescaled displacements on the 'good set' and of crack sets. The technical backbone is a refined jump transfer lemma in GSBD, a rotation-correction step that controls the region where the linearization is valid, and a stability-of-minimality argument that survives the limit. The paper also handles the broken-off pieces carefully via the bad set B(t), and Lemma 6.1 shows those pieces relax and do not pollute the energy balance. The energy convergence is proved for every time, not just a.e.\n\nThe soft spots are real but declared. The nonsimple-material assumption (second-gradient penalty with exponent beta in (2/3,1)) is load-bearing for the rotation decomposition and the cut-off construction; the paper says so plainly, and the stress-test note confirms it is an explicit hypothesis, not a hidden gap. It does mean the theorem does not apply to a purely first-gradient nonlinear energy, and removing it will likely need new rigidity ideas. Likewise, the 2D restriction comes from the piecewise Korn inequality, and the authors say higher dimensions need a new jump transfer lemma. The use of time-discrete approximate solutions rather than a true nonlinear evolution is also stated openly; this is more a feature than a bug, since the theorem is exactly about the discrete scheme, but it is worth flagging when citing.\n\nI did not verify every estimate line-by-line—that would take a GSBD specialist—but the proof structure is coherent and the cited tools are the right ones. The paper is honest about its limitations, which makes the result more credible, not less.\n\nFor anyone working on variational fracture or linearization, this is an important reference and deserves a serious referee. I would send it out without hesitation and expect it to be accepted, probably with minor revision.","headline":"Solid, genuinely new evolutionary linearization theorem; the nonsimple-material assumption is real but explicitly declared and doesn't undermine the result as stated.","tokens_in":45747,"tokens_out":2591,"would_cite":true,"duration_ms":28130,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["49J45","70G75","74B10","74B20","74G65","74R10"],"pacs":[],"model":"deepseek-v4-flash","headline":"For very stiff 2D brittle materials, time-discretized nonlinear quasistatic crack growth converges to the linear Griffith evolution, with no assumptions on crack geometry.","keywords":["brittle materials","variational fracture","free discontinuity problem","quasistatic fracture evolution","linearization","Γ-convergence","nonsimple materials","GSBD"],"falsifier":"A concrete test would be to construct, for some $\\beta\\in(2/3,1)$ and a sequence of time-discretized minimizers of the energy (2.4) with well-prepared initial data and uniformly bounded energies, a case where no subsequence of the rescaled displacements converges in measure on the good set, or where a cluster point violates the unilateral minimality inequality of Definition 2.1(iv). A numerical experiment with a branching crack in a stiff two-dimensional specimen could settle the question by exhibiting such a failure.","tokens_in":44844,"feed_emoji":"💥","tokens_out":8614,"duration_ms":85894,"temperature":0.7,"pith_summary":"This paper proves that in two dimensions, crack growth in a very stiff nonlinear elastic material is rigorously approximated by the standard linear Griffith model of brittle fracture. The authors show that rescaled deformations from a time-discretized nonlinear quasistatic evolution converge, as the stiffness tends to infinity, to a linear quasistatic fracture evolution: the limiting crack is the union of the jump sets of the limiting displacement, and the total energies converge at every time. This is the evolutionary counterpart of a known static linearization result, and it removes earlier restrictions that forced the crack to lie on a line segment or on finitely many regular arcs. The proof requires the nonlinear energy to penalize the second gradient of the deformation, a technical condition that disappears in the limit. The result matters because it justifies using linear theory to predict crack paths in stiff materials, even when the crack is tortuous and arbitrarily shaped.","feed_headline":"Nonlinear fracture converges to linear Griffith model","feed_subtitle":"Rigorous 2D proof drops crack-geometry assumptions: stiff-material crack paths match linear theory.","key_machinery":"The load-bearing machinery is a refined jump transfer lemma in the space GSBD of functions of bounded deformation, adapted from the linear fracture existence theory, together with a rotation decomposition into Caccioppoli pieces and a cut-off construction. The jump transfer lemma lets the authors move the unilateral global minimality of the nonlinear discrete minimizers to the linear limiting displacement: a competitor for the linear problem is lifted to a competitor for the nonlinear problem without creating jumps outside the accumulated crack, and errors are controlled by a quantitative almost-minimality estimate. The rotation decomposition, with cut-off sets $\\omega_\\varepsilon$, isolates on each piece the rigid rotation about which the deformation is linearized, while the cut-off removes the vanishingly small region where the rescaled gradient is too large for a Taylor expansion.","core_discovery":"The central claim is Theorem 2.2: given approximate nonlinear quasistatic evolutions $y_\\varepsilon$ built by time-discretized minimization of $E_\\varepsilon$, with well-prepared initial data, there is a linear quasistatic fracture evolution $(u(t),\\Gamma(t))$ such that $\\Gamma(t)=\\bigcup_{\\tau\\in I^t_\\infty} Ju(\\tau)$ and, up to a subsequence, $y_\\varepsilon\\rightsquigarrow u$, with $E_\\varepsilon(t)\\to E(t)$ for all $t\\in[0,1]$. The convergence is understood on the 'good set' attached to the Dirichlet boundary; broken-off pieces relax to rigid motions and carry no elastic energy in the limit. No a priori geometric restriction on the crack set is imposed.","pith_inferences":["Since the proof never uses a genuine continuous-time nonlinear solution and lets the time step and stiffness parameter go to zero independently, one likely extension is to full time-continuous nonlinear evolutions once existence and compactness are established.","The same machinery may transfer to anisotropic or heterogeneous crack growth, because the linear target enters only through the quadratic form $Q$ and the GSBD tools are model-independent.","For simulation practice, the result implies that linear-elastic fracture computations can be trusted to reproduce the crack path of a very stiff nonlinear material even when the crack is irregular; this is a testable prediction for stiff brittle sheets.","The Hessian term is currently load-bearing, so the theorem indicates where a counterexample to linearization without second-gradient regularization would have to come from; a first-gradient-only model needs a new rigidity mechanism."],"forward_implications":["No crack-shape assumptions: previous linearization results that pinned the crack to a line segment or to finitely many regular arcs are superseded in the two-dimensional nonsimple setting.","The limit crack is exactly the union of the jump sets of the linear displacement up to that time, so crack growth is determined by the linearized displacement field.","The total energy of the nonlinear evolution converges to the linear Griffith energy at every time, with separate convergence of elastic and crack contributions.","The limiting linear evolution satisfies irreversibility, unilateral global minimality, and the energy balance, so it is a quasistatic fracture solution in the sense of Definition 2.1.","The vanishing Hessian term shows the effective limit model is purely first-order linear Griffith elasticity, not a strain-gradient model."],"supporting_citations":[{"why":"Supplies the static gamma-convergence result for nonsimple brittle materials that defines the linear target energy and the well-prepared initial data.","marker":"[Fri20]"},{"why":"Provides existence of linear quasistatic fracture evolutions in GSBD and the jump transfer lemma that carries unilateral minimality to the limit.","marker":"[FS18]"},{"why":"Establishes the quasistatic evolution framework of irreversibility, unilateral global minimality, and energy balance that the limiting object must satisfy.","marker":"[FL03]"},{"why":"Gives the time-discrete scheme and a priori estimates for quasistatic crack growth in nonlinear elasticity used throughout the proof.","marker":"[DFT05]"},{"why":"Proves the planar piecewise Korn inequality used in the density and jump-transfer arguments.","marker":"[Fri18]"},{"why":"Supplies the density of regular functions in GSBD that reduces minimality to smooth competitors.","marker":"[Iur14]"},{"why":"Earlier linearization result restricted to straight crack paths, which the paper improves.","marker":"[NZ14]"},{"why":"Relaxes the crack-path restriction to finitely many regular arcs; the present work removes this condition.","marker":"[NT15]"}],"fun_headline_variants":["Stiff materials: nonlinear fracture converges to linear law","No crack-shape assumptions needed for linear fracture limit","Fracture linearization proven without geometric constraints","Quasistatic cracks: stiff limit recovers Griffith theory","Nonlinear to linear fracture: a geometry-free proof"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof depends on the nonlinear energy being nonsimple: it must contain a second-gradient (Hessian) penalty with exponent $\\beta\\in(2/3,1)$, because without that term the rotation decomposition and cut-off construction used to linearize break down; purely first-gradient nonlinear models are therefore not covered.","fun_headline_variants_meta":{"raw":{"variants":["Stiff materials: nonlinear fracture converges to linear law","No crack-shape assumptions needed for linear fracture limit","Fracture linearization proven without geometric constraints","Quasistatic cracks: stiff limit recovers Griffith theory","Nonlinear to linear fracture: a geometry-free proof"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000658,"raw_usage":{"total_tokens":2948,"prompt_tokens":820,"completion_tokens":2128,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":436,"completion_tokens_details":{"reasoning_tokens":2051}},"tokens_in":436,"tokens_out":2128,"duration_ms":15594,"temperature":1.0,"reasoning_tokens":2051,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T16:23:44.705925+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete test would be to construct, for some $\\beta\\in(2/3,1)$ and a sequence of time-discretized minimizers of the energy (2.4) with well-prepared initial data and uniformly bounded energies, a case where no subsequence of the rescaled displacements converges in measure on the good set, or where a cluster point violates the unilateral minimality inequality of Definition 2.1(iv). A numerical experiment with a branching crack in a stiff two-dimensional specimen could settle the question by exhibiting such a failure.","supporting_citations":[],"review_version":1}