{"id":"9219d47e-7aaf-42c4-adf9-99d48ef7ba22","arxiv_id":"2505.23348","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper proves a rigidity structure theorem and computes an explicit kernel projection for maps of bounded deviatoric deformation in dimension n≥3, providing the main tools for relaxation and homogenization in BD_dev.","lead":"This paper analyzes the space of functions of bounded deviatoric deformation, which appears in plasticity and fluid mechanics. It proves a rigidity theorem and constructs an explicit projection operator that together enable a blow-up strategy for energy relaxation and homogenization.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Density step in Theorem 5.3 lacks a canonical choice/control of the kernel part L_ε; boundedness of E_d alone does not imply L1 compactness.","rationale":"The reader identified the external De Philippis–Rindler theorem as the weakest assumption. That is a legitimate concern for the applicability of the result to blow-ups, but it does not affect the internal proof of Theorem 1.2 for maps already satisfying a constant-polar equation. The more load-bearing point is internal: the proof of Theorem 5.3 needs to pass from smooth mollifications to the general BD_dev map, and the passage relies on compactness of u_ε − L_ε. The paper gives no canonical choice of the kernel part L_ε and no argument that L_ε is uniformly controlled in L1; boundedness of E_d(u_ε − L_ε) = E_d u_ε is insufficient because the kernel of E_d is infinite in the relevant growth sense and contains unbounded sequences with zero E_d. All subsequent steps — convergence of the polynomial coefficients in Step one and the BV limits in Step two — are built on this asserted limit. The gap is likely repairable, since a fixed projection onto the finite-dimensional kernel would provide the missing compactness, but as written the proof is incomplete. For this reason I would condition acceptance on closing this density gap rather than rejecting the paper outright, as the core algebraic rigidity analysis appears coherent and the external tools cited are standard.","tokens_in":45582,"tokens_out":39668,"duration_ms":405638,"concrete_test":"Re-run the mollification argument of Section 5.1.3 with L_ε fixed as R_K[u_ε], where R_K is the kernel projection from Theorem 1.4 (or any fixed bounded projection onto Ker(E_d)). Verify whether the Poincaré inequality (2.10) then gives a uniform L1 bound for u_ε − L_ε and whether the coefficient limits for Q_ε and ψ_i^ε remain well-defined and independent of the chosen projection. If this repair works, the gap is cosmetic; if not, the rigidity theorem lacks a rigorous density step.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"In the proof of Theorem 5.3 (Section 5.1.3), the mollified maps u_ε = u ⋆ ρ_ε are decomposed using Proposition 5.4 as u_ε = aψ_1^ε(x·b) + bψ_2^ε(x·a) + Q_ε + L_ε, with L_ε ∈ Ker(E_d). The proof then asserts that because |E_d(u_ε − L_ε)|(A) is uniformly bounded and u_ε → u in L1, one has u_ε − L_ε → ū in L1. This implication is false as written: the kernel component L_ε is not specified and cannot be controlled by E_d alone. For example, taking L_ε = ε^{-1} y gives E_d L_ε = 0 while ∥L_ε∥_{L1} → ∞. The subsequent extraction of convergent polynomial coefficients Q_ε and of the BV limits of ψ_i^ε depends entirely on this compactness step. Since the decomposition of Proposition 5.4 is non-unique and no normalization or fixed projection onto Ker(E_d) is chosen, the density argument for the non-parallel rigidity theorem contains a real gap. If the gap cannot be closed by a canonical choice of L_ε, then Theorem 1.2 is not established by the given proof.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops the structural theory of the space BD_dev(Ω) of functions of bounded deviatoric deformation, defined through the trace-free symmetric gradient operator E_d u = sym Du - (div u)/n Id. The main results are: (i) an explicit fourth-order annihilator A for E_d whose wave cone is exactly {a⊙b - (a·b)/n Id : a,b∈R^n} (Propositions 3.2 and 3.6); (ii) a rigidity theorem (Theorem 1.2, proved as Theorems 5.3 and 5.11) describing every u∈BD_dev with E_d u equal to a constant polar matrix in the wave cone times a nonnegative measure as a sum of one-dimensional BV terms, a W^{1,1} part, a polynomial part solving an explicit deviatoric equation, and a Killing field in Ker(E_d); (iii) an explicit kernel projection operator R_K satisfying the Poincaré inequality (Theorem 1.4); and (iv) fine properties of E_d u, including |E_d u| ≪ H^{n-1}, the structure of the jump part, and a quasi-continuity property. The proof strategy is to mollify, solve the smooth PDE system derived in Lemma 5.1, and pass to the limit separately in the non-parallel and parallel cases, treating n=3 and n≥4 differently.","tokens_in":45826,"tokens_out":15854,"duration_ms":167908,"significance":"If the proof is repaired as suggested below, the paper would provide the BD_dev analogue of the classical BD rigidity and would supply the two main ingredients (rigidity with constant polar and an explicit kernel projection) needed for the iterative blow-up approach to relaxation and homogenization with integrands depending on u as well as E_d u. The explicit annihilator and the computation of its wave cone are clean and useful results, and the paper is honest about relying on the external De Philippis-Rindler theorem for the polar of the singular part. The paper also contains a clear statement of the open question whether |E_d u|(S_u \\ J_u)=0. These strengths are substantial. However, the central density step in Theorem 5.3 has a real compactness gap and the main theorem as stated contains a false regularity assertion; both are repairable within the manuscript's scope, but they block acceptance in the present form.","major_comments":[{"comment":"The step \"since sup_ε |E_d(u_ε−L_ε)|(A)<∞ we have u_ε−L_ε→ū in L1\" is not justified. The kernels L_ε are not specified and E_d does not control them: for example L_ε(y)=ε^{-1}y belongs to Ker(E_d) and has |E_d L_ε|=0 but ∥L_ε∥_{L1(A)}→∞. This is load-bearing because the subsequent extraction of the limits of Q_ε and of ψ_i^ε in Steps one and two depends on having a convergent subsequence of u_ε−L_ε. The gap can be repaired by fixing a bounded linear projection Π_A onto Ker(E_d), as already used in Proposition 2.4 or Proposition 2.6, and choosing L_ε=Π_A u_ε; the Poincaré inequality then gives a uniform L1 bound for u_ε−Π_A u_ε on compact subdomains, allowing the compactness theorem to be applied. This choice must be stated explicitly in the proof.","section":"Section 5.1.3 (proof of Theorem 5.3)"},{"comment":"The statement of the main theorem asserts ψ_1, ψ_2 ∈ C^∞(R), but the precise rigidity theorem (Theorem 5.3) proves only ψ_1, ψ_2 ∈ BV_loc(R), and this regularity is optimal. For instance u(x)=H(x·b)a belongs to BD_dev and satisfies E_d u=(a⊙b−(a·b)/n Id) μ with μ=H^{n-1} on {x·b=0}; one of the profile functions is a Heaviside function, not a C^∞ function. The statement of Theorem 1.2 should replace C^∞ by BV_loc; as printed, the central theorem is false.","section":"Theorem 1.2, Case 1"}],"minor_comments":[{"comment":"The labels ψ_1 and ψ_2 are interchanged relative to the statement (5.20): the proof writes w_ε=aψ^1_ε(x·b)+bψ^2_ε(x·a) and concludes w=aψ_1(x·b)+bψ_2(x·a), whereas (5.20) states ψ_1(x·a)b+ψ_2(x·b)a. The BV verification in Step three follows the proof's convention, so the statement and the proof must be reconciled.","section":"Section 5.1.3, Step two of Theorem 5.3"},{"comment":"The remark says \"our rigidity Theorems 5.3 and 5.11 are proven for n=2\", which contradicts both theorems and the surrounding discussion of C-ellipticity for n≥3; this is presumably intended to read \"proven for n≥3\" or \"not proven for n=2\".","section":"Remark 2.3"},{"comment":"The assertion that all coefficients of Q_ε are polynomial functions of η_ε, ϑ_ε, and v_ε is used to conclude Q_ε→Q, but the full coefficient dependence is not derived; Proposition 5.5 proves only the two displayed third-order coefficients. The authors should either justify this dependence systematically or restructure the argument so that Lemma 5.10 is applied directly to each component Q_ε·e_j.","section":"Section 5.1.3, Step one of Theorem 5.3"},{"comment":"The sentence \"for L^n-a.e. x∈ there exists a precise representative\" is missing the domain Ω; this is a typographical error.","section":"Section 4.1"},{"comment":"There are several minor typos, such as \"compacntess\" in Section 5.1.3 and \"wethern\" in Remark 2.3; a careful proofreading pass is needed.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The paper fits the journal's scope and the main ideas are sound. The density gap in Section 5.1.3 is real but repairable by inserting a fixed kernel projection; the C^∞ assertion in Theorem 1.2 is a straightforward but important correction. I do not see grounds for rejection, but the revision must address both items before the paper can be accepted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things you should know. First, the paper gives real new tools: a fourth-order annihilator for E_d, the wave cone computation, and an explicit kernel projection operator. Those are carefully done and will be useful beyond this paper. Second, the main rigidity theorem for the non-parallel case has a gap in the density argument, and I don't think Theorem 1.2 is established as written.\n\nThe gap is in Section 5.1.3. After mollification, the authors write u_ε = aψ1^ε(x·b) + bψ2^ε(x·a) + Q_ε + L_ε, L_ε ∈ Ker(E_d). They then claim that since |E_d(u_ε − L_ε)|(A) is bounded and u_ε → u in L1, one has u_ε − L_ε → ū in L1. That is false as stated. The kernel part L_ε is not specified, and boundedness of E_d gives no control on L1. Take L_ε = ε^-1 y. The decomposition from Proposition 5.4 is non-unique, so there is no canonical choice. All the subsequent convergence of Q_ε and ψ_i^ε depends on this compactness step. The paper later constructs a projection fixing Ker(E_d) in Theorem 1.4, but does not use it here. That would be the natural fix: replace L_ε by something like u_ε - R_K[u_ε] or fix a projection to remove the kernel ambiguity. As it stands, the proof of Theorem 5.3 does not go through.\n\nThe parallel case Theorem 5.11 is different. There the authors actually integrate against test functions and get explicit uniform bounds on P_j, Ψ etc. That proof is long but appears sound. The annihilator, wave cone, and kernel projection sections are independent of the flawed density step and look solid. I also noticed a likely typo in Remark 2.3: it says Theorems 5.3 and 5.11 are proven for n=2, but the paper assumes n≥3 for C-ellipticity; probably meant 'not proven for n=2'.\n\nWhere does this leave us? The paper deserves a serious referee, because the new tools are valuable and the gap is probably fixable with a known projection technique. But I wouldn't cite Theorem 1.2 in its current form. If you are working on relaxation or homogenization in BD_dev, read the annihilator and kernel projection sections now, and treat the non-parallel rigidity as a conjecture until the density step is repaired.","headline":"The annihilator, wave cone, and kernel projection are real contributions, but the density step in Theorem 5.3 has a genuine gap: L_ε is uncontrolled, and the asserted L1 compactness does not follow.","tokens_in":46339,"tokens_out":4075,"would_cite":false,"duration_ms":39734,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35B65","49J45","46E35","74C05"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that any $\\mathrm{BD}_{dev}$ map whose deviatoric strain has a constant polar direction in the wave cone splits into one-dimensional $BV$ profiles, a $W^{1,1}$ part, and a polynomial remainder.","keywords":["bounded deviatoric deformation","rigidity","wave cone","annihilator","kernel projection","blow-up","relaxation","homogenization"],"falsifier":"In $n=3$ with $a=b=e_1$, take the density $g$ from Lemma 5.13, $g=h(x_1)+\\sum_{j=2}^3 p_j(x_1)x_j+\\psi(x_1)(x_2^2+x_3^2)/2+\\varrho(x_2^2-x_3^2)$, and solve $\\mathcal{E}_d u = (e_1\\odot e_1 - \\mathrm{Id}/3)g$. Theorem 1.2 predicts the layered structure (5.51) with $F\\in BV_{\\mathrm{loc}}$; a solution whose non-polynomial part depends genuinely on both $x_2$ and $x_3$, or whose $F$ is not $BV$, would falsify the theorem. A simpler algebraic test is to compute $\\ker A[\\xi]$ from (3.7) and look for an element not of the form $a\\odot b-(a\\cdot b)/n\\,\\mathrm{Id}$.","tokens_in":45371,"feed_emoji":"📐","tokens_out":10658,"duration_ms":103485,"temperature":0.7,"pith_summary":"Functions of bounded deviatoric deformation, $\\mathrm{BD}_{dev}$, are vector fields whose deviatoric strain $\\mathcal{E}_d u = Eu - (\\mathrm{div}\\,u)/n\\,\\mathrm{Id}$ is a finite Radon measure; they arise in plasticity and fluid mechanics, where shear rather than compression is the controlled variable. This paper proves a rigidity theorem: if $\\mathcal{E}_d u$ equals a fixed trace-free symmetric matrix $a\\odot b - (a\\cdot b)/n\\,\\mathrm{Id}$ times a nonnegative measure, then $u$ must split into one-dimensional functions of $x\\cdot a$ and $x\\cdot b$ with $BV$ regularity, a $W^{1,1}$ piece, an explicitly described third-order polynomial $Q$, and a kernel element $L$. A companion theorem gives an explicit projection operator onto $\\mathrm{Ker}(\\mathcal{E}_d)$ and the resulting Poincar\\'e inequality. If true, the blow-up profiles needed for relaxation and homogenization of deviatoric energies are fully determined, including energies depending on $u$ itself.","feed_headline":"Constant shear-strain direction forces one-dimensional singular profiles","feed_subtitle":"Rigidity result pins down BD_dev blow-up profiles for plasticity and fluid-mechanics energies.","key_machinery":"The load-bearing object is the fourth-order annihilator $A$ for $\\mathcal{E}_d$, with symbol $A[\\xi]M=|\\xi|^2(M\\xi\\otimes\\xi+\\xi\\otimes M\\xi)-|\\xi|^4M-\\frac{\\xi^t M\\xi}{n-1}((n-2)\\xi\\otimes\\xi+|\\xi|^2\\mathrm{Id})$. Its kernel is exactly the wave cone $\\{a\\odot b-(a\\cdot b)/n\\,\\mathrm{Id}\\}$, which is the set of admissible constant polars. The proof then uses the curl-type identities of Lemma 5.1, obtained by differentiating the relation and applying Schwarz's theorem, to turn the constant-polar equation into wave equations for the density $g$; the remainder is a third-order polynomial term. A second mechanism is the explicit operator $R_K$ of Theorem 1.4, built from boundary integrals $s_K,A_K,\\gamma_K,b_K$, which fixes $\\mathrm{Ker}(\\mathcal{E}_d)$ and gives the Poincar\\'e estimate.","core_discovery":"The central claim, Theorem 1.2, is that for $n\\ge 3$ the wave cone of a fourth-order annihilator $A$ of $\\mathcal{E}_d$ is exactly the set $\\{a\\odot b - (a\\cdot b)/n\\,\\mathrm{Id}: a,b\\in\\mathbb{R}^n\\}$, and that any $u\\in \\mathrm{BD}_{dev}(K)$ with $\\mathcal{E}_d u = (a\\odot b - (a\\cdot b)/n\\,\\mathrm{Id})\\mu$ has one of two rigid forms. If $a,b$ are not parallel, $u(x)=\\psi_1(x\\cdot a)b+\\psi_2(x\\cdot b)a+Q(x)+L(x)$ with $\\psi_1,\\psi_2\\in BV_{\\mathrm{loc}}$, $Q$ a homogeneous third-order polynomial solving (1.6), and $L\\in\\mathrm{Ker}(\\mathcal{E}_d)$. If $b=\\lambda a$, $u$ has the layered structure (1.7) with $F\\in BV_{\\mathrm{loc}}$, $P_j,G\\in W^{1,1}_{\\mathrm{loc}}$ with derivatives in $BV_{\\mathrm{loc}}$; for $n\\ge 4$ the extra polynomial coefficient $\\varrho$ vanishes. The paper derives from the annihilator the fine structure $\\mathcal{E}_d u = e_d(u)\\mathcal{L}^n + ([u]\\odot\\nu_u - ([u]\\cdot\\nu_u)/n\\,\\mathrm{Id})\\mathcal{H}^{n-1}\\lrcorner J_u + a(x)\\odot b(x) - (a(x)\\cdot b(x))/n\\,\\mathrm{Id}\\,|\\mathcal{E}_d^c u|$, and Theorem 1.4 constructs an explicit operator $R_K$ fixing $\\mathrm{Ker}(\\mathcal{E}_d)$ with $\\|u-R_K[u]\\|_{L^1}\\le c|\\mathcal{E}_d u|(K)$.","pith_inferences":["The third-order polynomial remainder is a new feature with no $BD$ analogue; if it survives in the limit energies, homogenized shear-only models may contain cubic terms even when the underlying $BD$ model has none.","The $n=2$ failure of $C$-ellipticity suggests the rigidity statement may be false there; a two-dimensional counterexample with a singular part spread over a curve would test this.","The quasi-continuity obtained from $R_K$ is a natural stepping stone toward the paper's open conjecture $|\\mathcal{E}_d u|(S_u\\setminus J_u)=0$.","The same pipeline (annihilator, wave cone, constant-polar rigidity, kernel projection) could be applied to other trace-free first-order operators in continuum mechanics."],"forward_implications":["At singular (Cantor-type) points, every blow-up of a $\\mathrm{BD}_{dev}$ map has one of the two rigid forms in Theorem 1.2; this is the missing input for a full relaxation theorem on $\\mathrm{BD}_{dev}$.","Because $R_K$ is explicit and fixes the kernel, the iterated blow-up argument can compute the constant in front of the one-dimensional profile, as done for $BD$.","The annihilator gives $|\\mathcal{E}_d u|\\ll\\mathcal{H}^{n-1}$ and the jump part $[u]\\odot\\nu_u - ([u]\\cdot\\nu_u)/n\\,\\mathrm{Id}\\,\\mathcal{H}^{n-1}\\lrcorner J_u$, so the singular measure of $\\mathrm{BD}_{dev}$ maps is as concentrated as in $BD$.","Energies depending on $u$ as well as $\\mathcal{E}_d u$ can be handled, because the rigidity step does not require dropping dependence on $u$.","For $n\\ge 4$, the polynomial coefficient $\\varrho$ in the parallel case vanishes, so in high dimensions the layered structure is independent of the exceptional third-order polynomial."],"supporting_citations":[{"why":"Supplies the structure theorem for A-free measures used to place the polar of the singular part in the wave cone.","marker":"[16]"},{"why":"Provides the BD rigidity result with constant polar in the wave cone that the paper adapts to the deviatoric operator.","marker":"[17]"},{"why":"Establishes the iterative blow-up strategy for integral representation on BD that motivates Theorems 1.2 and 1.4.","marker":"[12]"},{"why":"Refines the double blow-up technique and kernel projection used to extend the approach to homogenization.","marker":"[13]"},{"why":"Provides the trace theory and C-ellipticity framework for the integration-by-parts and fine-property steps.","marker":"[11]"},{"why":"Gives the Poincar\\'e inequality for bounded linear kernel projections used to turn $R_K$ into the estimate (1.10).","marker":"[19]"},{"why":"Supplies the Poincar\\'e-Sobolev inequality for C-elliptic operators used in compactness.","marker":"[24]"},{"why":"Guarantees existence of an annihilator for the deviatoric operator, a starting point for the explicit fourth-order operator.","marker":"[33]"}],"fun_headline_variants":["Rigidity theorem classifies singular profiles in BD_dev","Wave cone identified for deviatoric deformation space","One-dimensional blow-ups forced by constant polar direction","New rigidity for elasticity with constant shear","Singular profiles in BD_dev pinned by annihilator"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument rests on the known theorem, not reproved here, that at almost every point of the singular part the polar of $\\mathcal{E}_d^s u$ lies in the wave cone; the whole blow-up strategy collapses if that theorem fails.","fun_headline_variants_meta":{"raw":{"variants":["Rigidity theorem classifies singular profiles in BD_dev","Wave cone identified for deviatoric deformation space","One-dimensional blow-ups forced by constant polar direction","New rigidity for elasticity with constant shear","Singular profiles in BD_dev pinned by annihilator"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000208,"raw_usage":{"total_tokens":1482,"prompt_tokens":1099,"completion_tokens":383,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":715,"completion_tokens_details":{"reasoning_tokens":310}},"tokens_in":715,"tokens_out":383,"duration_ms":4661,"temperature":1.0,"reasoning_tokens":310,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T12:48:40.888140+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In $n=3$ with $a=b=e_1$, take the density $g$ from Lemma 5.13, $g=h(x_1)+\\sum_{j=2}^3 p_j(x_1)x_j+\\psi(x_1)(x_2^2+x_3^2)/2+\\varrho(x_2^2-x_3^2)$, and solve $\\mathcal{E}_d u = (e_1\\odot e_1 - \\mathrm{Id}/3)g$. Theorem 1.2 predicts the layered structure (5.51) with $F\\in BV_{\\mathrm{loc}}$; a solution whose non-polynomial part depends genuinely on both $x_2$ and $x_3$, or whose $F$ is not $BV$, would falsify the theorem. A simpler algebraic test is to compute $\\ker A[\\xi]$ from (3.7) and look for an element not of the form $a\\odot b-(a\\cdot b)/n\\,\\mathrm{Id}$.","supporting_citations":[{"cited_title":"De Philippis and F","cited_arxiv_id":null,"evidence_quote":"Supplies the structure theorem for A-free measures used to place the polar of the singular part in the wave cone."},{"cited_title":"De Philippis and F","cited_arxiv_id":null,"evidence_quote":"Provides the BD rigidity result with constant polar in the wave cone that the paper adapts to the deviatoric operator."},{"cited_title":"On the integral representation of variational functionals on $BD$","cited_arxiv_id":"1907.11478","evidence_quote":"Establishes the iterative blow-up strategy for integral representation on BD that motivates Theorems 1.2 and 1.4."},{"cited_title":"Iterative blow-ups for maps with bounded $\\mathcal{A}$-variation: a refinement, with application to $\\mathrm{BD}$ and $\\mathrm{BV}$","cited_arxiv_id":"2504.02490","evidence_quote":"Refines the double blow-up technique and kernel projection used to extend the approach to homogenization."},{"cited_title":"Breit, L","cited_arxiv_id":null,"evidence_quote":"Provides the trace theory and C-ellipticity framework for the integration-by-parts and fine-property steps."},{"cited_title":"Diening and F","cited_arxiv_id":null,"evidence_quote":"Gives the Poincar\\'e inequality for bounded linear kernel projections used to turn $R_K$ into the estimate (1.10)."},{"cited_title":"Gmeineder and B","cited_arxiv_id":null,"evidence_quote":"Supplies the Poincar\\'e-Sobolev inequality for C-elliptic operators used in compactness."},{"cited_title":"Van Schaftingen","cited_arxiv_id":null,"evidence_quote":"Guarantees existence of an annihilator for the deviatoric operator, a starting point for the explicit fourth-order operator."}],"review_version":1}