{"id":"b0bf323a-07e1-4761-923e-0e88cb4c7e32","arxiv_id":"2411.17560","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Unobstructed deformations of complex parallelisable nilmanifolds are characterized by pseudo-free self-conjugate Lie algebras, classified up to dimension 20.","lead":"This paper proves that a compact complex parallelisable nilmanifold has unobstructed deformations exactly when its Lie algebra is pseudo-free and self-conjugate. It then classifies such Lie algebras up to dimension 20, finding 19 individual examples before dimension 20 and an infinite family in dimension 20.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Example 2.13 has a dimension inconsistency: the ideal a_mu as written has dimension 23, so g_mu would be 18-dimensional, not 20, and the infinite 20-dimensional family underpinning Corollary 1.3(iv) is not established.","rationale":"The paper's strongest claim is the full package: Theorem 1.1 plus the classification consequences, including the sharp cutoff 'finitely many up to dimension 19, infinitely many in dimension 20'. The infinite family in dimension 20 is a headline consequence and appears only through Example 2.13. That example contains an elementary dimension contradiction, so it is the most load-bearing concrete defect I can identify. The reader's weakest assumption also flags partially demonstrated computations U1(Vλ) and the under-explained bridge in Theorem 3.12; I agree these are secondary. Indeed, the bridge is probably repairable: because Φ_k vanishes for k > ν, the formal power series is polynomial, so evaluating at t = 1 after applying the gauge (3.7) yields a genuine Maurer–Cartan element and hence a Lie algebra homomorphism. The asserted computations in Section 2.3 are also checkable by the same I1/U1 method, whereas the dimension of a_mu can be checked using Table 3 alone. The verdict should therefore remain conditional: the main characterization may well be correct, but the dimension-20 classification and the corresponding homotopy-type corollary need correction or confirmation before they can be relied on.","tokens_in":85,"tokens_out":10817,"duration_ms":216506,"concrete_test":"Compute, with explicit highest weight vectors v1, v2 for the two copies of V(5,2) in f7, the dimension of the ideal generated by V(5,1), V(6,1), and the subspace spanned by v1, v2, and µ1 v1 + µ2 v2. If dim a_mu = 23 and dim(n_2,7/a_mu) = 18 as the displayed formula implies, the paper must be revised. Then determine whether the only way to obtain a 20-dimensional quotient is to take an ideal of dimension 21 such as V(5,1) ⊕ V(6,1) ⊕ V(5,2) ⊕ U_mu ⊕ V(4,3), verify by the I1/U1 procedures of Section 2.3 that this subspace is verbal, and check via Lemma 2.5 that different µ give pairwise non-isomorphic quotients.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The 20-dimensional family {g_mu} is the sole source of Corollary 1.3(iv) ('infinitely many complex homotopy types in dimension 20') and of the '1-parameter family of 20-dimensional Lie algebras' row in Table 1. But as written, a_mu = V(5,1) ⊕ V(6,1) ⊕ 2V(5,2) ⊕ U_mu ⊂ n_2,7 cannot define a 20-dimensional quotient. From Table 3, dim n_2,7 = 41, dim V(5,1) = 5, dim V(6,1) = 6, dim V(5,2) = 4, and U_mu ≅ V(5,2), so dim a_mu = 5 + 6 + 2·4 + 4 = 23, giving dim(n_2,7/a_mu) = 18, not 20. Moreover, Proposition 2.8 gives f7 = V(6,1) ⊕ 2V(5,2) ⊕ 2V(4,3) for two generators, so only two copies of V(5,2) exist; '2V(5,2) ⊕ U_mu' would require a third copy. A 20-dimensional quotient would need an ideal of dimension 21, for example V(5,1) ⊕ V(6,1) ⊕ V(5,2) ⊕ U_mu ⊕ V(4,3) (dimension 5 + 6 + 4 + 4 + 2 = 21), but the text neither states nor proves that this is the intended ideal. Until this arithmetic is corrected and verified, Corollary 1.3(iv), Theorem 1.2(iii), and the dimension-20 row of Table 1 are unsupported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a criterion for unobstructed deformations of compact complex parallelisable nilmanifolds in terms of pseudo-free Lie algebras and a reality condition, and it provides a partial classification of pseudo-free Lie algebras up to dimension 20, claiming that infinite families first appear in dimension 20. The main theorem is proved by reducing unobstructedness to the existence of Lie algebra homomorphisms extending linear maps from an abelian quotient, and the classification uses representation-theoretic computations of verbal ideals in free Lie algebras.","tokens_in":16990,"tokens_out":15534,"duration_ms":126895,"significance":"If the main theorem and the classification hold, the paper gives a substantial structural result: unobstructedness is governed by two purely algebraic properties, and the classification up to dimension 19 with an infinite family in dimension 20 is a striking new phenomenon. The paper's strength is its clean algebraic reformulation in Lemma 3.11 and its connection of deformation theory to verbal ideals, which could be of independent interest. However, several load-bearing computations and the dimension-20 family are not yet rigorously established, so the significance is conditional on a careful revision.","major_comments":[{"comment":"The ideal aµ = V(5,1) ⊕ V(6,1) ⊕ 2V(5,2) ⊕ Uµ ⊂ n2,7 has dimension 5+6+8+4 = 23 (using Table 3), so dim(n2,7/aµ) = 41−23 = 18, not 20 as claimed. Moreover, Proposition 2.8 shows that f7 contains only two copies of V(5,2), so the expression '2V(5,2) ⊕ Uµ' would require a third copy, which does not exist. The family {gµ} is therefore not defined as written, and the assertions in Theorem 1.2(iii), Table 1, and Corollary 1.3(iv) about a 1-parameter family of 20-dimensional pseudo-free Lie algebras are unsupported. This also undermines the abstract's claim that infinite families first appear in dimension 20.","section":"Example 2.13 / Table 1 / Corollary 1.3"},{"comment":"The proof of the key equivalence 'µ is unobstructed iff there exists a Lie algebra homomorphism Φ: g → g extending µ' is not spelled out. From a Kuranishi power series Φ(t) = Σ Φk(t) with Φ1 = µ, one can evaluate at a fixed t to obtain an element Φ ∈ g∗⊗g satisfying the Maurer–Cartan equation, but the restriction of Φ to V = g/[g,g] is not necessarily µ, because the higher-order terms Φk (k ≥ 2) may have nonzero values on V. The condition that Φk has no ∂-closed summands does not imply that Φk vanishes on V. The text's sentence 'the Φ1-part of Φ, regarded as a linear map g→g, is simply given by its restriction to V' conflates the first-order part of the series with the total evaluated map. This is a load-bearing step for Theorem 1.1.","section":"Theorem 3.12"},{"comment":"The classification depends on several asserted computations of the form U1(Vλ) and U2(Vλ), for example U1(V(3,1)) = f5, U1(V(4,1)) = V(5,1) ⊕ V(4,2), U1(V(3,2)) = V(4,2) ⊕ V(3,3), U2(V(4,1)) = f7, and U1(V(5,1) ⊕ V(4,2)) = V(6,1) ⊕ 2V(5,2) ⊕ V(4,3). These are justified only by the phrase 'This can be done as in Example 2.17' or are simply listed. Since these identities determine the list in Table 1 and the dimension cutoff 20, the authors should provide full proofs or a reproducible computation; otherwise the classification is not verifiable.","section":"Proposition 2.19 / Theorem 2.20"}],"minor_comments":[{"comment":"The phrase '20-dimensional' in the statement 'Then {gµ = n2,7/aµ}µ∈P1_C defines a family of 20-dimensional pseudo-free Lie algebras' should be corrected once the dimension of the quotient is computed correctly.","section":"Example 2.13"},{"comment":"The statement that the family {gµ} 'can occur at least for µ ∈ P1_Q' conflicts with the reality condition; only parameters with µ ∈ P1_R (and rational to admit a lattice) satisfy gµ ≅ gµ, so the parameter set for unobstructed manifolds is P1_Q ∩ P1_R, not all of P1_Q.","section":"Corollary 3.14"},{"comment":"The sentence 'The only 7-step pseudo-free Lie algebras relevant to our classification are the 17-dimensional Lie algebra n2,7/a from Example 2.18 and the Lie algebras gµ from Example 2.13' needs to be re-evaluated if the dimension of gµ is not 20.","section":"Proof of Theorem 2.20"},{"comment":"The notation g ∼= g for the complex conjugate Lie algebra is used in the introduction and in Theorem 1.1 but is only defined later in Section 3.1; a brief definition at first use would improve readability.","section":"Introduction / Section 3.1"}],"recommendation":"major_revision","confidential_remarks":"The dimension error in Example 2.13 is the kind of issue that could be fixed by the authors, but as it stands the paper's headline claim about dimension 20 is not established. I suggest the editor ask for a careful revision with full computational appendices and a repaired proof of the bridge in Theorem 3.12. The paper is within scope for math.DG, and the algebraic framework is promising."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the good news: Theorem 1.1 is a genuine advance. The characterization of unobstructed deformations of complex parallelisable nilmanifolds in terms of pseudo-freeness plus the reality condition g ≅ g is exactly the kind of clean statement the field needed, and Lemma 3.11 (Maurer–Cartan solutions = Lie algebra homomorphisms) is a neat observation that makes the bridge between deformation theory and verbal ideals look natural. The connection to verbal ideals and Zhuravlev's theorem is well-motivated, and the classification up to dimension 19 is a substantial piece of work.\n\nBut the paper has a load-bearing arithmetic problem in Example 2.13. As written, the ideal a_mu = V(5,1) ⊕ V(6,1) ⊕ 2V(5,2) ⊕ U_mu ⊂ n2,7 has dimension 5+6+8+4=23 (using dim V(5,1)=5, dim V(6,1)=6, dim V(5,2)=4 from Table 3), so n2,7/a_mu would be 18-dimensional, not 20. Moreover, U_mu is defined as a subrepresentation of the two copies of V(5,2) in f7, so writing 2V(5,2) ⊕ U_mu is double-counting; you cannot have an extra independent copy. The text gives no other way to read this. As a consequence, the one-parameter family {g_mu} that underpins Corollary 1.3(iv) and the dimension-20 row of Table 1 is not established. This is not a minor typo—it removes the only stated source of the infinite family in dimension 20.\n\nThe rest of the soft spots are less severe. The classification relies on a number of assertions like U1(V(4,1)) = V(5,1) ⊕ V(4,2) with 'this can be done as in Example 2.17', which are plausible but not fully demonstrated in the text. A referee will want those computations either verified or tabulated in an appendix. The step in Theorem 3.12 from existence of a formal power series to a single homomorphism extension is also compressed; the evaluation at t=1 deserves a few more sentences.\n\nThe main theorem, though, appears to hold. The logic of the proof is sound, and the reduction to pseudo-freeness is convincing. The paper is honestly written and the authors acknowledge the partial nature of the classification. I'm not claiming the dimension-20 result is false—it may well be true with a different ideal or a different construction—but as it stands the paper doesn't support it.\n\nNet: this deserves a serious referee, exactly because the main theorem matters and the mistake is a concrete, fixable error rather than a deep conceptual problem. I'd send it to a competent referee and ask them to focus on Example 2.13 and the computational identities. If the authors correct the family (or remove it and adjust Corollary 1.3), the paper becomes solid.","headline":"Main theorem is a real advance, but the dimension-20 family in Example 2.13 doesn't add up; fix that and the paper is solid.","tokens_in":90,"tokens_out":4299,"would_cite":true,"duration_ms":99626,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["32G05","17B01","17B30","32M10"],"pacs":[],"model":"deepseek-v4-flash","headline":"Two algebraic conditions decide unobstructed nilmanifold deformations","keywords":["complex parallelisable nilmanifold","unobstructed deformations","Kuranishi space","pseudo-free Lie algebra","verbal ideal","free nilpotent Lie algebra","variety of Lie algebras","complex homotopy type"],"falsifier":"Compute, by an independent linear-algebra calculation, the space $I_1(V(4,1))$ inside the degree-6 component of the free Lie algebra on two generators; if it is not the 8-dimensional representation $V(5,1) \\oplus V(4,2)$, the classification table collapses. Alternatively, for the 17-dimensional algebra $\\mathfrak{n}_{3,4}/V(3,1)$, compute the quadratic term of the Kuranishi series for a generic infinitesimal deformation; nonzero quadratic obstruction would contradict Theorem 1.1.","tokens_in":16412,"feed_emoji":"🧮","tokens_out":10683,"duration_ms":93761,"temperature":0.7,"pith_summary":"The paper identifies exactly when a compact complex parallelisable nilmanifold has unobstructed deformations: when its associated nilpotent complex Lie algebra is pseudo-free and is isomorphic to its complex conjugate. Pseudo-free means the algebra is the free algebra of some variety of Lie algebras, equivalently a free Lie algebra cut down by a verbal ideal. If the criterion is right, unobstructedness is a purely algebraic, checkable property, and the paper pushes the resulting classification far enough to show a sharp dimension cutoff: finitely many such manifolds up to dimension 19, infinitely many in dimension 20. This matters because unobstructedness is usually hard to test and no general vanishing theorem of that kind is available in this non-Kähler setting.","feed_headline":"Two algebraic conditions decide unobstructed nilmanifold deformations","feed_subtitle":"A sharp new criterion shows finitely many such manifolds below dimension 20, then infinitely many.","key_machinery":"The carrying object is the notion of a pseudo-free Lie algebra: a nilpotent Lie algebra that is the free Lie algebra of some variety of Lie algebras, equivalently a quotient of a free Lie algebra by a verbal (fully invariant) ideal. The key bridge is Lemma 3.11, which shows that a vector-valued 1-form $\\Phi \\in \\mathfrak{g}^* \\otimes \\mathfrak{g}$ solves the Maurer–Cartan equation exactly when $\\Phi$ is a Lie algebra homomorphism; unobstructedness then becomes the statement that every linear map from $\\mathfrak{g}/[\\mathfrak{g},\\mathfrak{g}]$ to $\\mathfrak{g}$ extends to an endomorphism, which is precisely pseudo-freeness plus the reality condition that the conjugate algebra is isomorphic to the original. The classification side is carried by the decomposition of each homogeneous component of the free Lie algebra into irreducible $GL(V)$-representations, with verbal ideals determined by their highest-weight pieces and propagated by derivations via the formula $\\mathfrak{h}_{n+k} = U_k(\\mathfrak{h}_n)$.","core_discovery":"The central claim is Theorem 1.1: a compact complex parallelisable nilmanifold $X$ with associated Lie algebra $\\mathfrak{g}$ has unobstructed deformations if and only if $\\mathfrak{g}$ is pseudo-free and $\\mathfrak{g} \\cong \\bar{\\mathfrak{g}}$ as complex Lie algebras. The only-if direction follows because every infinitesimal deformation must extend to a genuine homomorphism of the Lie algebra, and the existence of such extensions for all linear maps $V \\to \\mathfrak{g}$ is exactly the defining property of a pseudo-free Lie algebra; the reality condition is forced by the need for a homomorphism $\\mathfrak{g} \\to \\bar{\\mathfrak{g}}$ fixing the generators. The paper then classifies non-abelian pseudo-free Lie algebras: for nilpotency index at most 3 they are freely nilpotent, for index at most 5 they form a short explicit list, and up to dimension 20 there are 19 individual algebras plus one 1-parameter family, with infinite families beginning exactly in dimension 20. Translating back to geometry gives finitely many complex homotopy types of unobstructed complex parallelisable nilmanifolds up to dimension 19 and infinitely many in dimension 20.","pith_inferences":["Beyond the paper, the same two-condition criterion could be turned into an algorithm: for any nilpotent complex Lie algebra given by structure constants, pseudo-freeness and the conjugacy condition are decidable by linear algebra, so unobstructedness of the corresponding manifold would be checkable by computer.","The dimension-20 cutoff is probably not an accident of the tables: infinite families appear exactly where a homogeneous component of the free Lie algebra contains an irreducible representation with multiplicity at least 2, suggesting the finiteness phenomenon is controlled by representation-theoretic multiplicities in general.","If the paper's expectation is right that no analogous statement holds for general nilmanifolds with left-invariant complex structure, the verbal-ideal criterion marks a clean boundary between the parallelisable case and the wider class; finding a left-invariant non-parallelisable example whose obstructions are not governed by pseudo-freeness would test that boundary.","The 1-parameter family in dimension 20 with rational parameter values gives a natural test bed for explicit deformation computations: one could try to write down the full Kuranishi family for $g_\\mu$ and see how the obstructions or holomorphic quantities vary with $\\mu$."],"forward_implications":["Unobstructedness of these manifolds is reduced to a finite, checkable algebraic computation on the Lie algebra's structure constants.","For nilpotency index at most 3, unobstructedness forces the Lie algebra to be freely nilpotent; at most 5, the possible Lie algebras are exactly those listed in Table 2.","Up to dimension 19 there are finitely many complex homotopy types of unobstructed complex parallelisable nilmanifolds, and in dimension 20 there are infinitely many.","Examples with unobstructed deformations whose Lie algebra is not freely nilpotent first appear in dimension 10, with a 5-step Lie algebra, and the smallest 4-step example has dimension 17.","Every freely nilpotent Lie algebra satisfying the reality condition $\\mathfrak{g} \\cong \\bar{\\mathfrak{g}}$ produces an unobstructed manifold, so the classical Iwasawa-type examples fit the criterion as a special case."],"supporting_citations":[{"why":"Supplies the Kuranishi space construction and the obstruction map that defines unobstructedness.","marker":"[Kur62]"},{"why":"Shows the deformation problem of these nilmanifolds reduces to a finite-dimensional differential graded Lie algebra and that the Kuranishi series terminates.","marker":"[Rol11]"},{"why":"Provides the theorem that a verbal ideal is generated from its relations by derivations, the engine for the classification computations.","marker":"[Zhu97]"},{"why":"Supplies the theory of varieties, verbal ideals, free Lie algebras, and the dimension formulas used throughout Section 2.","marker":"[Bah21]"},{"why":"Provides the Schur module decomposition into irreducible GL(V)-representations used to detect verbal ideals.","marker":"[Ful96]"},{"why":"Gives the lattice criterion that turns the classified complex Lie algebras into compact complex parallelisable nilmanifolds.","marker":"[Mal51]"}],"fun_headline_variants":["Pseudo-free Lie algebras unlock nilmanifold deformations","Unobstructed nilmanifolds: finite to 19, infinite at 20","Two conditions solve nilmanifold deformation question","Verbal ideals reveal nilmanifold deformation boundary"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument hinges on two things not fully spelled out in the text: the listed decomposition computations are correct, and a formal power series solution of the Maurer–Cartan equation always comes from a single homomorphism extension.","fun_headline_variants_meta":{"raw":{"variants":["Pseudo-free Lie algebras unlock nilmanifold deformations","Unobstructed nilmanifolds: finite to 19, infinite at 20","Two conditions solve nilmanifold deformation question","Verbal ideals reveal nilmanifold deformation boundary"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000261,"raw_usage":{"total_tokens":1571,"prompt_tokens":899,"completion_tokens":672,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":515,"completion_tokens_details":{"reasoning_tokens":600}},"tokens_in":515,"tokens_out":672,"duration_ms":6765,"temperature":1.0,"reasoning_tokens":600,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T12:02:25.421026+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute, by an independent linear-algebra calculation, the space $I_1(V(4,1))$ inside the degree-6 component of the free Lie algebra on two generators; if it is not the 8-dimensional representation $V(5,1) \\oplus V(4,2)$, the classification table collapses. Alternatively, for the 17-dimensional algebra $\\mathfrak{n}_{3,4}/V(3,1)$, compute the quadratic term of the Kuranishi series for a generic infinitesimal deformation; nonzero quadratic obstruction would contradict Theorem 1.1.","supporting_citations":[],"review_version":1}