{"id":"df3da2c6-8f0e-4d11-b0ec-941a2847d66c","arxiv_id":"2508.13342","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":2,"one_line_summary":"A new elliptic reformulation of surface Stokes lets any finite element spaces be used, avoiding the discrete inf-sup condition while preserving optimal convergence.","lead":"The authors rewrite the surface Stokes equations as a system of two Laplace equations, which no longer requires the discrete inf-sup condition. This lets any polynomial degree pair be used for velocity and pressure while retaining proven optimal error rates.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central claim holds only for Stokes data F in H^{-1}_t(divΓ)×H^1_# (and H_t(divΓ)×H^2_# for optimal L2); for merely L^2 pressure data the reformulated data G falls outside V^{-1,-1}, so the method as analyzed does not cover the full original Stokes problem.","rationale":"The reader's conditional verdict is justified. The central construction—the elliptic reformulation (3.3) and the Schatz-type discrete analysis—is internally coherent: the equivalence chain in §3.1 is valid for sufficiently regular data, Theorem 3.2 supplies continuous well-posedness, and Lemmas 4.5–4.8 together with Theorem 4.9 give the advertised quasi-optimality and L2 rates for h≤h0. The numerical experiments are consistent with the theory. The main load-bearing qualification is data regularity: the discrete unknowns live in H1_t×H1_#, so the pressure is sought in H1; consequently the reformulated data G must lie in V^{-1,-1}, and in V^{0,0} for the optimal L2 theorem. The paper explicitly discloses this in Remarks 4.3 and 4.13, and my stress test constructs a valid Stokes solution with p∈L2\\H1 whose reformulated datum is outside V^{-1,-1}, showing the restriction is not vacuous. The omitted proof of Lemma 3.5 is a smaller gap: it is used in (4.17) to obtain the duality estimate; the statement appears true by elliptic regularity applied to the adjoint system, but a rigorous proof should be added. The C9 formula in Lemma 4.8 also contains a denominator inconsistency (cA,1 versus cA,1^2) that should be corrected. None of this invalidates the central theorem in its regular-data setting; it means the title's unqualified claim must be read with the disclosed data-regularity caveat. Hence the verdict remains conditional.","tokens_in":25755,"tokens_out":31343,"duration_ms":326608,"concrete_test":"On S^2, choose p∈L^2_#\\H^1_#, for example p = Σ_k k^{-3/2} Y_k for spherical harmonics Y_k, set u=0 and F=(∇p,0). Verify that (0,p) solves the original Stokes problem (2.11) with data F, while G=D(F)=(∇p,-Δp) has second component in H^{-2}\\setminus H^{-1}. This confirms that G∉V^{-1,-1}, so the discrete RHS G[V_{h,t}] is not well-defined for V_{h,t}∈V^{1,1}; the data-regularity restriction in Remarks 4.3 and 4.13 is therefore essential rather than cosmetic.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The discrete analysis applies to the elliptic reformulation (3.3) with G∈V^{-1,-1}, not to the original Stokes problem in its minimal data space V^{-1,0}. For Stokes-generated data G=D(f,f)=(f,-divΓf-ΔΓf), membership in V^{-1,-1} requires f∈H^{-1}_t(divΓ) and f∈H^1_# (Remark 4.3); the optimal L2 theorem additionally requires G∈V^{0,0}, i.e., f∈H_t(divΓ) and f∈H^2_# (Remark 4.13). This is one to two orders above the minimal setting. Take F=(∇p,0) with p∈L^2_#\\H^1_#; then (0,p) is a valid Stokes solution, but the reformulated second datum is g=-Δp∈H^{-2}\\setminus H^{-1}, so G∉V^{-1,-1} and the discrete right-hand side G[V_{h,t}] is not defined for V_{h,t}∈V^{1,1}. Thus the title's unqualified claim is narrower than the theorem. A secondary rigor gap is that Lemma 3.5 (adjoint H2 regularity) is asserted without proof and is needed for the L2 duality estimate (4.15) and the threshold h0; it appears true but should be supplied, along with the C9 denominator typo in Lemma 4.8.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a new formulation of the surface Stokes problem on a closed C^3 hypersurface, rewriting the saddle-point system (2.11) as a nonsymmetric indefinite elliptic system (3.3) of two Laplacians via the identity (2.12). The authors prove well-posedness of the continuous problem in V^{1,1} for data in V^{-1,-1}, derive a Garding inequality, and analyze a lifted parametric FEM with a weak tangentiality penalty. Following Schatz's approach, they prove discrete well-posedness for h <= h0 (depending only on the Weingarten map), a quasi-best approximation property in a mesh-dependent H^1 norm, and optimal L^2 error estimates for both velocity and pressure, for arbitrary polynomial degree pairs. Numerical experiments on a torus confirm the predicted convergence rates.","tokens_in":26085,"tokens_out":27949,"duration_ms":240592,"significance":"The paper's main contribution is a discretization of surface Stokes that provably avoids any discrete inf-sup condition: the elliptic reformulation (3.3) is well-posed with respect to the V^{1,1} norm, and the finite element analysis in Sections 4.7-4.11 delivers quasi-optimality and optimal L^2 rates for any velocity-pressure pair. The analysis is largely self-contained: the crucial identity (2.12) is proved in Lemma 2.4, well-posedness is connected to the standard Stokes problem, and the Schatz-style duality argument is carried out in detail. The numerical experiments are reproducible (code on GitHub) and show optimal orders, including for pairs like P1/P2 and P1/P3 that are not inf-sup stable for the original formulation. The main caveat is the data regularity required for the reformulation to be equivalent to Stokes; this should be prominently stated.","major_comments":[{"comment":"The paper presents the method as a reformulation of the surface Stokes equations, but the equivalence and the discrete analysis hold only for Stokes data F=(f,f) with f in H^{-1}_t(divΓ) and f in H^1_# (for well-posedness) and f in H_t(divΓ), f in H^2_# (for the optimal L^2 estimate). Indeed, the elliptic system (3.3) is analyzed for G in V^{-1,-1}, whereas the data map D in (3.2) sends the minimal Stokes data space V^{-1,0} only to V^{-1,-2} (Lemma 3.1). For pressure data f merely in L^2_#, the second component g = -divΓ f - ΔΓ f lies in H^{-2}_# but not H^{-1}_#, so G[V_{h,t}] in (4.11) is not defined for V_{h,t} in V^{1,1}. The authors acknowledge this in Remark 4.3, but the abstract and introduction claim an unqualified reformulation of the surface Stokes equations. I recommend stating the data regularity in the abstract and introduction, and adding a brief discussion of how the method relates to, or could be extended for, the minimal-data setting.","section":"Abstract; §3.3 (Theorem 3.2); §4 (Remark 4.3)"},{"comment":"Lemma 3.5 (higher regularity of the adjoint problem) is asserted without proof, with the note that the proof is similar to that of (3.9). This lemma is load-bearing: it provides the V^{2,2} regularity of the adjoint solution used in Lemma 4.7 to obtain the factor h in the L^2 error estimate, which in turn enters the definition of h0 in Lemma 4.8 and the final L^2 estimate in Theorem 4.11. Since the adjoint problem (3.14) has different signs and coupling than the forward problem, the proof is not literally identical to the forward case. Please supply the proof or a detailed sketch.","section":"§3.4 (Lemma 3.5); used in §4.8 (Lemma 4.7)"}],"minor_comments":[{"comment":"The constant C9 is stated as C9 = 2 max{1,C_A^2}/c_{A,1}, but the proof (just before (4.19)) yields C9 = 2 max{1,C_A^2}/c_{A,1}^2. Please correct the statement.","section":"§4.9 (Lemma 4.8)"},{"comment":"The stated continuity constant C_A = 1 + C_{P,t} max{1, ||R||_{0,∞}} is not consistent with the proof: the four terms give a bound of 2 + C_{P,t}(1 + ||R||_{0,∞}). Since the exact value is not used qualitatively, this is a minor correction, but the stated value should be valid.","section":"§3.5 (Proposition 3.6)"},{"comment":"The derivation gives kh(I_h v, I_h v)^{1/2} = (1/τ)||n·I_h v|| ≤ (C/τ) h^2 ||v||_2. To conclude the stated O(h) bound with a constant independent of τ, one needs h ≲ τ (or the constant must be allowed to depend on τ). Please state this explicitly or record the τ-dependence of C7 and the subsequent constants.","section":"§4.6 (Lemma 4.5)"},{"comment":"Minor typos: 'parameteric' should be 'parametric'; the abstract has a missing space in 'C3'; consider using \\bar{w} consistently for the pressure in (3.3)–(3.4) to avoid ambiguity.","section":"§6 (Conclusions)"}],"recommendation":"major_revision","confidential_remarks":"The core analysis is sound and the numerical results support the theory; the main reservation is the scope of the claims relative to the data regularity. The companion paper [BNS25] is cited for background, but the essential identity (2.12) is proved here, so the dependency is not load-bearing. The paper fits the scope of math.NA."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a genuinely new idea with the proofs mostly in place. The authors rewrite the surface Stokes problem as an indefinite elliptic system driven by two Laplacians and show the bilinear form satisfies a Garding inequality, so any conforming finite element pair works without a discrete inf-sup condition once the mesh resolves the Ricci curvature. That is a real advance over the older stabilization/perturbation tricks, and the numerical experiments back it up, including equal-order pairs that are not inf-sup stable for the primitive formulation.\n\nWhat is new and solid: the equivalence is demonstrated through an identity (Lemma 2.4) that is proved in the paper, not merely cited, so the companion [BNS25] is not load-bearing. The continuous well-posedness and the discrete analysis follow the Schatz–Aubin–Nitsche route cleanly. The numerics are honest: manufactured solutions, a nontrivial torus with sign-changing curvature, and rates that match the theory. The code is public.\n\nThe soft spots are real but not fatal. The main one, which the authors disclose in Remarks 4.3 and 4.13 but the abstract underplays, is that the reformulated elliptic problem is equivalent to the original Stokes problem only when the data lies in H^{-1}_t(divΓ) × H^1_#; the optimal L2 result needs H_t(divΓ) × H^2_#. This is one to two orders above the minimal V^{-1,0} data space. The stress-test example is correct: F=(∇p,0) with p in L^2_# but not H^1_# gives a perfectly valid Stokes solution, yet the reformulated data falls outside V^{-1,-1}, so the method as analyzed does not even have a well-defined right-hand side. The title should therefore be read with that regularity qualifier in mind. I would ask the authors to state this limitation more prominently, perhaps in the abstract, but it does not undermine the core construction.\n\nTwo smaller issues: Lemma 3.5 (adjoint H2 regularity) is asserted without proof and is needed for the duality estimate and the mesh threshold; it looks true, but it should be proved rather than hand-waved. And in Lemma 4.8 the constant C9 is printed with c_{A,1} in the denominator; the proof gives c_{A,1}^2. That is a typo, not a substantive error.\n\nBottom line: this paper deserves a serious referee and likely publication after the regularity qualification is made explicit and the missing adjoint-regularity proof is supplied. I would bring it to a reading group and would certainly cite it if I worked on surface Stokes.","headline":"A genuinely new elliptic reformulation of surface Stokes that sidesteps the discrete inf-sup condition, with clean analysis and convincing numerics — but the equivalence to the original Stokes problem holds only for data one to two orders more regular than the minimal setting.","tokens_in":26644,"tokens_out":4455,"would_cite":true,"duration_ms":46805,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35J47","35J50","35M30","35Q35","58J05","65N12","65N15","65N30"],"pacs":[],"model":"deepseek-v4-flash","headline":"A two-Laplacian reformulation of the surface Stokes problem eliminates the discrete inf-sup condition for arbitrary element pairs.","keywords":["surface Stokes","elliptic reformulation","inf-sup condition","parametric FEM","Weingarten map","Ricci curvature","Gårding inequality","error estimates"],"falsifier":"On the torus example of Section 5 with the same manufactured velocity but a pressure data $f$ chosen in $L^2_\\# \\setminus H^1_\\#$, run the $P1/P1$ discretization of the reformulated system: if the discrete pressure still converges in $L^2$ to the true pressure, the regularity restriction in Remark 4.3 is unnecessary; if it fails to converge or converges to a different limit, the restriction is real.","tokens_in":25532,"feed_emoji":"🌊","tokens_out":9603,"duration_ms":84998,"temperature":0.7,"pith_summary":"This paper shows that the tangential Stokes problem on a closed $C^{3}$ hypersurface can be rewritten exactly as a nonsymmetric, indefinite elliptic system in which both the velocity and the pressure satisfy Laplace-type equations. For this two-Laplacian reformulation, a lifted parametric finite element method is well-posed for any combination of velocity and pressure polynomial degrees, provided the mesh size h is below a threshold h0 that depends only on the surface's Weingarten map. The paper proves quasi-best approximation in a mesh-dependent $H^{1}$ norm and optimal-order $L^{2}$ error estimates for both unknowns. The result matters because the discrete inf-sup condition, which normally forces specific element pairs or stabilization, is completely circumvented.","feed_headline":"Two Laplacians replace the inf-sup condition in surface Stokes","feed_subtitle":"New elliptic reformulation lets any velocity-pressure element pair run stably with optimal error rates.","key_machinery":"The load-bearing identity is the decomposition of the Bochner Laplacian, $a(v, \\nabla q) = (\\mathrm{div}\\, v, \\Delta q) - (Rv, \\nabla q)$ for $v \\in H^1_t$ and $q \\in H^2_\\#$, with $R = \\mathrm{tr}(W) W - W^2$ the Ricci curvature map. It converts the Stokes constraint into a pressure equation, producing the two-Laplacian system (3.3). The second engine is the argument for discrete well-posedness of indefinite Galerkin problems: the Gårding inequality leaves a compact term, which is absorbed using a conditional $L^2$ error estimate obtained by a duality argument; this yields a threshold $h_0$ that shrinks as the deviation of $R$ from the identity grows.","core_discovery":"The central claim is that the primitive-variable surface Stokes system (2.11) is equivalent to the elliptic system (3.3) on the space $V^{1,1} = H^1_t \\times H^1_\\#$, where the pressure is one order more regular than in the classical formulation. Testing the momentum equation with gradients of $H^2$ test functions and using the decomposition $a(v, \\nabla q) = (\\mathrm{div}\\, v, \\Delta q) - (Rv, \\nabla q)$ turns the divergence constraint $\\mathrm{div}\\, u = f$ into a second Laplace equation for the pressure, yielding two coupled Laplace problems perturbed only by lower-order terms involving the Ricci curvature map $R = \\mathrm{tr}(W) W - W^2$. The induced bilinear form is continuous, satisfies a Gårding inequality, and is an isomorphism on $V^{1,1}$. Consequently, for $h \\le h_0$ with $h_0$ depending only on the Weingarten map, the lifted parametric FEM with arbitrary finite element spaces is uniquely solvable and satisfies $E_1^{h,1}[W,W_h] \\le C_9 \\inf_{V_h} E_1^{h,1}[W,V_h]$ and $\\|W - W_{h,t}\\|_{0,0} + h E_1^{h,1}[W,W_h] \\le C_{11} h^2 \\|G\\|_{0,0}$.","pith_inferences":["Because the pressure is sought in $H^1_\\#$ instead of $L^2_\\#$, the method trades the inf-sup condition for one extra order of data regularity; a natural next test is whether a pressure-stabilized variant could recover convergence for minimal-regularity data.","The two-Laplacian structure suggests block-diagonal preconditioners built from scalar Laplace-Beltrami solvers, which may be noticeably cheaper than standard Stokes preconditioners.","The same reformulation strategy could apply to other vector Laplace operators (for instance Hodge Laplacians) whenever a Ricci-type curvature identity relates the differential operators.","On evolving surfaces, discretizing the reformulated system in space and time might avoid inf-sup issues in the tangential Navier-Stokes setting, though this is not addressed in the paper."],"forward_implications":["Any velocity-pressure pair of polynomial degrees, equal or disparate, can be used without stabilizers or inf-sup-stable element choices, simplifying surface Stokes codes.","The method achieves optimal-order convergence in both variables: $\\|W - W_{h,t}\\|_{0,0} + h E_1^{h,1}[W,W_h] = O(h^2)$ for sufficiently smooth data, with no dependence of the convergence order on the element pair.","The same analysis covers the physically relevant Stokes model with the perturbed surface diffusion operator $\\alpha I - \\Delta_S$, $\\alpha > 0$, in two dimensions.","The smallness condition $h \\le h_0$ is tied to resolving the deviation of the Ricci curvature map from the identity on the tangent space, not to resolving principal curvatures.","The reformulation provides new high-regularity results for the Stokes solution when data lie in $H^{-1}_t(\\mathrm{div}\\,\\Gamma) \\times H^1_\\#$."],"supporting_citations":[{"why":"Proves the decomposition of the Bochner Laplacian (Lemma 2.1) and supplies the elliptic regularity results on C^3 manifolds used throughout the analysis.","marker":"[BNS25]"},{"why":"Establishes the vector Laplacian identity $P\\,\\mathrm{div}\\,\\Gamma \\nabla^T_\\Gamma v = \\nabla_\\Gamma \\mathrm{div}\\,\\Gamma v + R v$ and the well-posedness of the surface Stokes problem.","marker":"[JOR18]"},{"why":"Provides the duality-based argument showing discrete well-posedness of Galerkin methods for indefinite problems.","marker":"[Sch74]"},{"why":"Introduces lifted parametric finite elements for surface PDEs, the discretization family analyzed in this paper.","marker":"[Dzi88]"},{"why":"Supplies the lifted interpolation and norm equivalence framework for surface finite element methods.","marker":"[DE13b]"},{"why":"Supplies the duality argument used in the conditional L^2 error estimate.","marker":"[Nit70]"},{"why":"Supplies the interpolation operators with the H^2 approximation estimates needed for the quasi-optimality and error bounds.","marker":"[CD15]"}],"fun_headline_variants":["Two Laplacians, zero inf-sup for surface Stokes","Surface Stokes: elliptic reformulation skips inf-sup","Bypass inf-sup with a dual-Laplacian Stokes reformulation","No inf-sup condition: surface Stokes as two Laplacians"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the data $F = (f, f)$ for the original Stokes problem lies in $H^{-1}_t(\\mathrm{div}\\,\\Gamma) \\times H^1_\\#$, one order more regular in both the force divergence and the pressure than the minimal Stokes data space $H^{-1}_t \\times L^2_\\#$; if the pressure data is only $L^2$, the reformulated system's $H^1$ pressure may fail to recover the true pressure.","fun_headline_variants_meta":{"raw":{"variants":["Two Laplacians, zero inf-sup for surface Stokes","Surface Stokes: elliptic reformulation skips inf-sup","Bypass inf-sup with a dual-Laplacian Stokes reformulation","No inf-sup condition: surface Stokes as two Laplacians"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000321,"raw_usage":{"total_tokens":1828,"prompt_tokens":984,"completion_tokens":844,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":600,"completion_tokens_details":{"reasoning_tokens":770}},"tokens_in":600,"tokens_out":844,"duration_ms":7958,"temperature":1.0,"reasoning_tokens":770,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T17:13:42.778494+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On the torus example of Section 5 with the same manufactured velocity but a pressure data $f$ chosen in $L^2_\\# \\setminus H^1_\\#$, run the $P1/P1$ discretization of the reformulated system: if the discrete pressure still converges in $L^2$ to the true pressure, the regularity restriction in Remark 4.3 is unnecessary; if it fails to converge or converges to a different limit, the restriction is real.","supporting_citations":[],"review_version":2}