{"id":"cac154cd-81aa-4cdb-b52f-1303eacc43df","arxiv_id":"2412.08028","paper_version":4,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper computes an explicit local formula for the spectral Einstein functional of the Dirac operator with torsion on closed even-dimensional spin manifolds.","lead":"This mathematics paper derives an explicit formula for the spectral Einstein functional of the Dirac operator with torsion on closed even-dimensional spin manifolds. The result gives a local integral expression involving the Einstein tensor, squared torsion terms, and covariant derivatives of torsion.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The torsion coefficients in Theorem 3.3 rest entirely on imported Lemma 2.1 and the Lichnerowicz formula; a direct flat-torsion or 4D cross-check is needed before the formula can be accepted.","rationale":"The paper's own algebra is internally coherent: the T=0 limit of Theorem 3.3 reduces to the standard Einstein functional, and the internal cancellation that combines B1 and B2 into the final G(v,w) term works. The reader identified the right weak spot: Lemma 2.1 and Eq. (2.3) are imported without proof or independent verification. My reading confirms that every torsion term in the final formula depends on exactly those imported inputs, so a single sign or factor error there would propagate into the headline result without breaking the torsion-free test. I disagree with the reader's secondary claim that the metric functional (3.4) has a normalization error: under the residue convention (3.1), the leading symbol of D_T^{-n} is |ξ|^{-n} id, and tracing c(v)c(w) gives exactly the stated constant. The proposed flat-torsion computation is the most direct way to test the torsion sector in isolation: it exercises the T_a and E terms without curvature complications, and the linear-T variant exercises the T_ab term that produces ∇T in the final formula. Until such a check is done, the result should remain conditional rather than accepted.","tokens_in":41494,"tokens_out":31297,"duration_ms":297541,"concrete_test":"Run a computer-algebra check on a flat torus T^{2m} with constant skew 3-form T and constant v,w. In this case ∇T=0, dT=0, curvature zero, so Theorem 3.3 predicts B_DT = 2^m Vol(S^{2m-1}) ∫ [-9/2 ||T||^2 g(v,w) + 9/2 Σ_{j,l} T(v,e_j,e_l)T(w,e_j,e_l)] dVol. Compute the left-hand side directly from the constant-coefficient full symbol of D_T = i c(ξ) + (3/2) T_cliff using the residue integral (3.1), for m=2 and m=3. If the coefficients disagree, Lemma 2.1 and/or (2.3) is wrong. If they agree, repeat with T(x)=T_0+Σ x_b L_b (still flat, so ∇T≠0) to verify the T_ab / ∇T term; a mismatch there would localize the error to Eq. (2.14) or (2.8).","verdict_should_be":"UNCHANGED","load_bearing_attack":"Central formula Theorem 3.3 is a long algebraic identity whose torsion-dependent terms all flow from two imported ingredients: Lemma 2.1 (from [1]) for the homogeneous symbol expansion of Δ_{T,E}^{-m}, and the torsion Lichnerowicz formula Eq. (2.2)-(2.3) (from [13]) giving E = 3/2 dT + 1/4s - 3/4||T||^2. In particular, the final coefficients -9/2||T||^2 g(v,w), +9/2Σ T(v,e_j,e_l)T(w,e_j,e_l), +3/2Σ∇_{e_a}T(e_a,v,w), and +3Σ T(v,∇w,e_j) are obtained by substituting T_a, T_ab and E into Lemma 2.1. If any of those substitutions carries a wrong sign or factor — for example if the -3/4||T||^2 in (2.3) or the T_a term in (2.10) is off — every torsion term in Theorem 3.3 changes, while the torsion-free limit T=0 would still hold. That is precisely why the imported lemma is load-bearing. The paper reports no independent check of Lemma 2.1 and does not compare with the known 4-dimensional computation in [13]. The reader's additional suspicion about the normalization of A_DT is, however, not supported: with σ_1(D_T)=√-1 c(ξ), the leading symbol of D_T^{-n} is |ξ|^{-n} id, which yields (3.4) exactly. So the central open risk is confined to the torsion sector, not the metric functional.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper computes the Wodzicki-residue functionals A_DT and B_DT for the Dirac operator with torsion D_T on a closed even-dimensional spin manifold. Theorem 3.3 states that A_DT = -2^m (2π^m/Γ(m)) ∫_M g(v,w) dVol and that B_DT = Wres(c(v)(c(w)D_T + D_T c(w))D_T^{-n+1}) is given by an explicit local integral involving the Einstein tensor G(v,w), the torsion norm squared, the bilinear term Σ T(v,e_j,e_l)T(w,e_j,e_l), the covariant derivative ∇T, and a term T(v,∇^L w,·). The proof splits B_DT into B1 = Wres(c(v)c(w)D_T^{-n+2}) and B2 = Wres(c(v)D_T c(w)D_T D_T^{-n}), relies on the symbol expansion of Δ_{T,E}^{-m} imported from [1] and on the torsion Lichnerowicz formula (2.2)–(2.3) from [13], and supplies an appendix of Clifford trace identities.","tokens_in":41820,"tokens_out":30877,"duration_ms":264592,"significance":"If Theorem 3.3 is correct, the paper gives a complete local description of the spectral Einstein functional for Dirac operators with torsion in arbitrary even dimension, generalizing the torsion-free result of [1] and matching the expected Einstein term when T=0. The computation is systematic and includes machine-checkable-style trace identities; there are no fitted parameters and no circular assumption of the final formula. The main risk is that every torsion-dependent coefficient in Theorem 3.3 is obtained by substituting (2.3), (2.7), (2.8) into the imported Lemma 2.1, so an incorrect sign or normalization in those inputs would change the final formula while preserving the torsion-free limit.","major_comments":[{"comment":"The torsion coefficients in Theorem 3.3 all flow from Lemma 2.1, imported from [1], and from the torsion Lichnerowicz formula (2.2)–(2.3), imported from [13]. The substitution of (2.3), (2.7), and (2.8) into Lemma 2.1 is stated as Lemma 2.2 without a detailed check. Because a wrong sign or factor in T_a, T_ab, or E would change every torsion term in the final formula while leaving the T=0 limit unchanged, this is load-bearing. I request either a full derivation of Lemma 2.2 from the definitions of D_T and Δ_{T,E}, or an independent cross-check, for example comparison with the 4-dimensional computation in [13] and a flat-torsion test.","section":"§2, Lemmas 2.1–2.2 and Theorem 3.3"},{"comment":"The dT contribution to B2 is discarded using the identity Σ_{i<j<k<t} tr(c(v)c(e_f)c(w)c(e_f)c(e_i)c(e_j)c(e_k)c(e_t)) = 0, stated in (3.58) without proof. This identity is not an immediate consequence of Lemma A.1 for general m, since products of six Clifford generators can have nonvanishing trace when m>2. Because this is the only place where the dT part of E enters the computation, a proof or a correct derivation is needed before the final formula can be regarded as fully verified.","section":"§3, Part II-3-H, Eq. (3.58)"},{"comment":"The subprincipal symbol σ0(D_T) is imported from [13] with no proof, and its connection-term sign is not reconciled with the convention in (2.4), where ∇_{∂_a} = ∂_a + (1/4) Σ ⟨∇^L_{∂_a} e_s,e_t⟩ c(e_s)c(e_t). The curvature contribution to σ0(AB) in Eq. (3.26), and hence the metric part of BDT, depends on this sign. The authors should state explicitly how Lemma 3.4 follows from the conventions of Section 2, or derive σ0(D_T) directly.","section":"§3, Lemma 3.4 and Eq. (2.4)"}],"minor_comments":[{"comment":"The displayed normalization `-2m 2π^m / Γ(m)` is ambiguous in the text; if it is intended to be `-2^m 2π^m / Γ(m)`, the superscripts should be typeset clearly, since the same expression later appears as `2 m 2πm / Γ(m)`.","section":"Theorem 3.3 and Eqs. (3.4), (3.5)"},{"comment":"The text says one needs to compute ∫ tr[σ_{-2m}(P1 D_T^{-2m})], but the quantity is AB D_T^{-2m}, not P1 D_T^{-2m}; this is a typo that should be corrected.","section":"§3, Part II, before Eq. (3.25)"},{"comment":"The term 3/2 Σ ∂T/∂x_j c(v)c(dx_j)c(ew)c(e_f)c(e_α)c(e_β) appears to contain a typo: `c(ew)` should presumably be `c(w)`, with the ∂w term handled separately in the following term.","section":"§3, Eq. (3.29)"},{"comment":"There are several typographical issues, including \"there form\" for \"three-form\" and the title \"Dirac operator wit h torsion\"; the paper should be carefully proofread.","section":"Throughout"},{"comment":"Lemmas 3.4 and 3.5 are stated without proof or precise reference to [13]; since Lemma 3.6 depends on their signs and normalizations, a short derivation or explicit pointer would improve readability.","section":"§3, Lemmas 3.4–3.6"}],"recommendation":"major_revision","confidential_remarks":"The paper is a long algebraic computation extending prior work by the same group and by Dabrowski et al. The novelty is incremental but within the journal's scope. The central formula is plausible and the torsion-free limit is correct, but the reliability of the torsion sector depends on imported symbol formulas and on one unproved trace cancellation. I recommend major revision rather than rejection, asking for a verification of Lemma 2.2 and a proof of Eq. (3.58)."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper likely gets the right closed formula for the spectral Einstein functional with torsion in even dimensions. The B1+B2 decomposition is coherent, the Clifford trace appendix is detailed, and the T=0 limit reproduces the standard -1/6 G(v,w) term. But the companion metric functional (3.4) has a normalization error: it states -2m 2π^m/Γ(m), whereas the paper's own conventions (tr[id]=2^m, unnormalized sphere measure) give -2^m 2π^m/Γ(m). For m=2 they agree; for m≠2 they don't. The stress-test note's claim that the leading symbol yields (3.4) exactly is mistaken. This looks like a typo (2m instead of 2^m), and it does not threaten the B_DT result.\n\nWhat's new: the explicit even-dimensional no-boundary formula for B_DT with torsion. It's a natural extension of the program in [1,13,23], and the new content is the assembly. The assembly is done carefully: the m-dependent terms cancel in the final sum, which is a good sign, and the known zero-torsion limit checks out.\n\nSoft spots: besides the (3.4) typo, the torsion part of B_DT rests entirely on imported Lemma 2.1 and the torsion Lichnerowicz formula in (2.2)-(2.3). No independent check is reported, not even the 4D case or a flat-torsion limit. A wrong coefficient in those imports would change every torsion term while leaving T=0 intact. The paper also has many typos and minor notational slips that don't affect the math but make verification harder.\n\nWho it's for: people working on spectral actions and noncommutative residues with torsion. It's a solid technical contribution once the metric functional typo is fixed and a cross-check is added. I'd send it to a referee—the main result deserves scrutiny—but I'd tell the referee to verify (3.4) and ask the authors to check a known limit.","headline":"Main Einstein functional formula is plausible and internally consistent, but (3.4) has a normalization typo, and the torsion sector wants a cross-check.","tokens_in":678,"tokens_out":1581,"would_cite":false,"duration_ms":92622,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53C27","58J40"],"pacs":[],"model":"deepseek-v4-flash","headline":"Explicit local formula found for the spectral Einstein functional with torsion.","keywords":["Dirac operator with torsion","spectral Einstein functional","noncommutative residue","spin manifold","torsion","symbol expansion","Lichnerowicz formula"],"falsifier":"On a flat $2m$-dimensional torus with constant skew-symmetric torsion $T$ and parallel unit vector fields $v,w$, the claimed formula reduces to $\\frac{2^m 2\\pi^m}{\\Gamma(m)}\\int_M \\bigl(-\\frac{9}{2}\\|T\\|^2 g(v,w)+\\frac{9}{2}\\sum_{j,l}T(v,e_j,e_l)T(w,e_j,e_l)\\bigr)$. An independent Fourier-mode calculation of $\\mathrm{Wres}(c(v)(c(w)D_T+D_T c(w))D_T^{-n+1})$ on this torus would either reproduce that number or contradict Theorem 3.3.","tokens_in":41286,"feed_emoji":"📐","tokens_out":12723,"duration_ms":100824,"temperature":0.7,"pith_summary":"The paper aims to compute, on a closed even-dimensional spin manifold, the spectral Einstein functional built from the Dirac operator with skew torsion $D_T$: the noncommutative residue of $c(v)(c(w)D_T + D_T c(w))D_T^{1-n}$. It claims that this functional is exactly a local integral whose integrand is a combination of the Einstein tensor contracted with $v,w$, a torsion-norm term, a torsion cross term, a covariant-divergence term of the torsion, and a term involving the Levi-Civita derivative of $w$. If the computation is right, spectral geometry with torsion supplies a concrete, computable analogue of the gravitational action that reduces to the torsion-free case when $T=0$.","feed_headline":"Explicit formula for the spectral Einstein functional with torsion","feed_subtitle":"On even-dimensional closed spin manifolds the residue equals a local integral of Ricci, torsion, and torsion derivatives.","key_machinery":"The central machinery is the noncommutative residue, a trace over the unit cosphere bundle of the $-n$-order symbol of a pseudodifferential operator; the spectral Einstein functional is defined through it. The computation is carried by the asymptotic symbol expansion of $\\Delta_{T,E}^{-m}$ with $\\Delta_{T,E}=D_T^2$, imported from references [1] and [13], together with the torsion Lichnerowicz formula $D_T^2 = -g^{ab}(\\nabla_{\\partial_a}\\nabla_{\\partial_b}-\\nabla_{\\nabla^L_{\\partial_a}\\partial_b}) + E$, where $E=\\frac{3}{2}dT+\\frac{1}{4}s-\\frac{3}{4}\\|T\\|^2$. The Clifford trace identities in the appendix reduce products of Clifford symbols to metric contractions, and the composition formula for symbols organizes the product $c(v)D_T c(w)D_T D_T^{-n}$ into six symbol-order terms that are integrated over the unit sphere.","core_discovery":"On an $n=2m$ dimensional closed spin manifold with torsion, the paper's Theorem 3.3 asserts $A_{D_T} = \\mathrm{Wres}(c(v)c(w)D_T^{-n}) = -\\frac{2^m 2\\pi^m}{\\Gamma(m)}\\int_M g(v,w)\\,d\\mathrm{Vol}_M$ and $B_{D_T} = \\mathrm{Wres}(c(v)(c(w)D_T + D_T c(w))D_T^{-n+1}) = \\frac{2^m 2\\pi^m}{\\Gamma(m)}\\int_M \\bigl(-\\frac{1}{6}G(v,w) - \\frac{9}{2}\\|T\\|^2 g(v,w) + \\frac{9}{2}\\sum_{j,l} T(v,e_j,e_l)T(w,e_j,e_l) + \\frac{3}{2}\\sum_a \\nabla_{e_a}(T)(e_a,v,w) + 3\\sum_j T(v,\\nabla^L_{e_j}w,e_j)\\bigr)\\,d\\mathrm{Vol}_M$, where $G(v,w)=\\mathrm{Ric}(v,w)-\\frac{1}{2}s\\,g(v,w)$. The proof splits $B_{D_T}$ into $B_1=\\mathrm{Wres}(c(v)c(w)D_T^{-n+2})$ and $B_2=\\mathrm{Wres}(c(v)D_T c(w)D_T D_T^{-n})$, computes each $-2m$-order symbol using the pseudodifferential composition formula and Clifford trace identities, and sums all contributions to reach the displayed formula.","pith_inferences":["A direct test the paper does not run: compute both sides on a flat torus with constant torsion via Fourier modes; if the numbers disagree, the imported symbol expansion is the first place to recheck.","The derivative term $\\sum_a \\nabla_{e_a}(T)(e_a,v,w)$ and the connection term $\\sum_j T(v,\\nabla^L_{e_j}w,e_j)$ are not obviously symmetric in $v,w$, so $B_{D_T}$ may be an asymmetric bilinear form; the paper does not discuss this.","The same symbol-expansion strategy should extend to the Hodge–Dirac operator with torsion or to manifolds with boundary, where the analogous spectral Einstein functional is not yet computed.","Because the torsion enters through three visibly different channels (pointwise norm, cross term, and derivative terms), the formula suggests that spectral measurements of the residue could in principle separate intrinsic torsion from its gradient, an inversion problem the paper leaves untouched."],"forward_implications":["The functional is local: the residue depends only on the metric, the torsion tensor, and its first covariant derivatives at each point, not on global topological data.","Setting $T=0$ recovers the torsion-free spectral Einstein functional, so $B_{D_T}$ specializes to a multiple of $\\int_M G(v,w)\\,d\\mathrm{Vol}_M$ in the ordinary Dirac case.","On a flat manifold with constant torsion and parallel $v,w$, only the torsion-norm and torsion-cross terms survive, so the formula predicts a purely torsion-driven residue.","The same computation yields the metric functional $A_{D_T}$ as a universal multiple of $\\int_M g(v,w)\\,d\\mathrm{Vol}_M$, independent of the torsion.","Definition 3.7 turns the computed expression into a candidate spectral Einstein functional with torsion on every closed even-dimensional spin manifold."],"supporting_citations":[{"why":"Defines the metric and Einstein functionals and supplies the symbol expansion of $(\\Delta_{T,E})^{-m}$ used as the starting point.","marker":"[1]"},{"why":"Supplies the torsion Lichnerowicz formula $D_T^2 = -g^{ab}(\\nabla_a\\nabla_b-\\nabla^L\\nabla)+E$ with $E=\\frac32 dT+\\frac14 s-\\frac34\\|T\\|^2$ and the symbols of $D_T$.","marker":"[13]"},{"why":"Provides Definition 3.1 of the spectral Einstein functional in the form used in the paper.","marker":"[23]"},{"why":"Gives the direct proof that the noncommutative residue of $D^{-2}$ recovers the Einstein–Hilbert action, the baseline this computation extends.","marker":"[6]"},{"why":"Establishes the noncommutative residue as a trace functional on pseudodifferential operators, the object being computed.","marker":"[16]"}],"fun_headline_variants":["Exact spectral Einstein residue for torsion Dirac","Torsion's trace: closed-form spectral Einstein functional","Dirac torsion yields clean spectral Einstein integral","Spectral Einstein functional: torsion-encoded exact formula"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the symbol expansion of $(D_T^2)^{-m}$ and the torsion Lichnerowicz formula with $E=\\frac{3}{2}dT+\\frac{1}{4}s-\\frac{3}{4}\\|T\\|^2$, both imported from prior references without re-derivation, are correct; if either is wrong, the final integrand changes.","fun_headline_variants_meta":{"raw":{"variants":["Exact spectral Einstein residue for torsion Dirac","Torsion's trace: closed-form spectral Einstein functional","Dirac torsion yields clean spectral Einstein integral","Spectral Einstein functional: torsion-encoded exact formula"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000235,"raw_usage":{"total_tokens":1466,"prompt_tokens":874,"completion_tokens":592,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":490,"completion_tokens_details":{"reasoning_tokens":532}},"tokens_in":490,"tokens_out":592,"duration_ms":6636,"temperature":1.0,"reasoning_tokens":532,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T18:22:57.986424+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On a flat $2m$-dimensional torus with constant skew-symmetric torsion $T$ and parallel unit vector fields $v,w$, the claimed formula reduces to $\\frac{2^m 2\\pi^m}{\\Gamma(m)}\\int_M \\bigl(-\\frac{9}{2}\\|T\\|^2 g(v,w)+\\frac{9}{2}\\sum_{j,l}T(v,e_j,e_l)T(w,e_j,e_l)\\bigr)$. An independent Fourier-mode calculation of $\\mathrm{Wres}(c(v)(c(w)D_T+D_T c(w))D_T^{-n+1})$ on this torus would either reproduce that number or contradict Theorem 3.3.","supporting_citations":[{"cited_title":"Dabrowski., A","cited_arxiv_id":null,"evidence_quote":"Defines the metric and Einstein functionals and supplies the symbol expansion of $(\\Delta_{T,E})^{-m}$ used as the starting point."},{"cited_title":"W ang, Y","cited_arxiv_id":null,"evidence_quote":"Supplies the torsion Lichnerowicz formula $D_T^2 = -g^{ab}(\\nabla_a\\nabla_b-\\nabla^L\\nabla)+E$ with $E=\\frac32 dT+\\frac14 s-\\frac34\\|T\\|^2$ and the symbols of $D_T$."},{"cited_title":"Kastler.: The Dirac Operator and Gravitation","cited_arxiv_id":null,"evidence_quote":"Gives the direct proof that the noncommutative residue of $D^{-2}$ recovers the Einstein–Hilbert action, the baseline this computation extends."},{"cited_title":"W odzicki","cited_arxiv_id":null,"evidence_quote":"Establishes the noncommutative residue as a trace functional on pseudodifferential operators, the object being computed."}],"review_version":1}