{"id":"89a5b1f6-b993-4627-a30a-bd71df41ad43","arxiv_id":"2507.06084","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A constrained Floer homology is defined by restricting the symplectic area functional to loops of zero mean Hamiltonian, and is shown to be well-defined under a strengthened Weinstein condition.","lead":"This paper defines a new Floer homology for periodic orbits on Liouville boundaries, built from the symplectic area functional constrained to loops with zero mean Hamiltonian. It proves the moduli spaces are Fredholm and compact only under a strong 'gradient type' condition, and defers the advertised product structure to a sequel.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The displayed linearization in Lemma 1.21 omits the derivative of the non-local factor chi; Theorem 1.28's Fredholm comparison therefore treats chi as frozen, leaving the actual constrained operator's Fredholm property unproved.","rationale":"The reader's weakest_assumption was the gradient-type condition. I partially agree: that condition is honestly flagged and it is what controls compactness, so it is not a hidden flaw. The more decisive issue is internal: the linearization of the constrained flow is mis-stated. The term dχ is not a local vector-field derivative; it involves integrals over the loop, so the citation to [AD14] does not cover it. If the operator ξ↦dχ_u(ξ)∇H|_u is non-compact, the index comparison with RFH may fail; if it is compact, the paper still needs to prove it. This is a proof gap in the main Fredholm theorem, not a disagreement with standard theory. The paper is otherwise honest: it flags the gradient-type assumption, the deferred product structure, and the reliance on 'usual arguments'. Given that the gap is likely repairable, I would not change the CONDITIONAL verdict, but the revision should supply the missing χ-derivative analysis and the compactness argument.","tokens_in":17395,"tokens_out":11693,"duration_ms":145220,"concrete_test":"Recompute dF_u directly from the definition by expanding F(exp_u(εξ)) for ξ∈L^p(R,u^*TH) and isolate the coefficient of ε; verify whether it equals (11). If the missing term dχ_u(ξ)∇H|u is nonzero, test its compactness: take u_0(s,t)=(r_0+s,t) on the cylindrical end and ξ_n(s,t)=φ_n(s)ψ(t), where φ_n is a bounded oscillating sequence in W^{1,p}_δ(R) and ψ is a fixed smooth loop; compute ∥dχ(ξ_n)∇H∥_{L^p_δ}. If no subsequence converges, the omitted term is non-compact and the reduction in Theorem 1.28 fails as written.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The constrained flow operator is F(u)=∂su+J(u)∂tu+χ(u)∇H|u. For a variation ξ(s)∈T_{u(s)}H, the derivative of the last summand should be dχ_u(ξ)∇H|u + χ(u)∇_ξ(∇H|u), but formula (11) displays only ∇_ξ(J(u)∂tu+χ(u)∇H|u), with no dχ term. The proof of Theorem 1.28 then compares with dG(u,χ(u)); in the Rabinowitz operator the multiplier is an independent variable, so dG contains no dχ. Thus the argument establishes Fredholmness of the frozen-chi operator, not of the true dF. If dχ is a compact perturbation in the δ-weighted Sobolev spaces the conclusion could be salvaged, but no compactness argument for this non-local term appears. Since Corollary 1.29 and the finite-dimensionality of M(x−,x+) rest on Theorem 1.28, the central claim is not established as written. Separately, Lemma 2.7 uses the global identity H=e^r−1, which is only stated near Σ; under the gradient-type condition ∇H=Λ it does follow globally, but that implication is not recorded.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a new Floer-type homology, called mean value constrained Floer homology (CFH), for periodic Reeb orbits on the boundary of a Liouville domain. The construction replaces the Lagrange multiplier of Rabinowitz Floer homology by an intrinsic constraint: one restricts the symplectic area functional to the hypersurface H = h^{-1}(0), where h is the Hamiltonian mean value functional. The main technical results are a Fredholm comparison theorem (Theorem 1.28) asserting that the linearized gradient-flow operator of the constrained functional is Fredholm if and only if the corresponding Rabinowitz operator is, with the same index, and a compactness theorem (Theorem 2.10) for the resulting moduli spaces, conditional on an L-infinity bound for the non-local constraint factor chi. That bound (Lemma 2.2) is proved under an additional geometric hypothesis, namely that the Liouville vector field is of gradient type (nabla H = Lambda). The final section sketches the Morse-Bott construction of the homology and states independence of auxiliary data.","tokens_in":17505,"tokens_out":12408,"duration_ms":139434,"significance":"The idea of avoiding the Lagrange multiplier is conceptually appealing and could lead to an intrinsic product structure on Floer homology for contact boundaries, a point the author explicitly postpones to a sequel. The paper does prove a genuine a priori bound (Lemma 2.1) from first principles, and the Fredholm reduction to Rabinowitz Floer homology, if completed, would be a useful structural result. However, the correctness of the central construction is currently conditional on closing two technical gaps: one in the linearization formula for the constrained operator, and one in the global use of the cylindrical-end form of the Hamiltonian. For this reason the paper should not be accepted as is, but a careful revision could make the main claims sound.","major_comments":[{"comment":"The displayed linearization in Lemma 1.21 as written is not the derivative of the actual constrained flow operator. Since F(u) = ∂_s u + J(u)∂_t u + χ(u)∇H|_u and χ(u) is a non-local function of the whole loop u(s,·), the variation of the last summand along ξ contains the term dχ_u(ξ)∇H|_u, which is absent from (11) if the symbol ∇_ξ is interpreted in the standard way for a vector field depending on u through χ(u). More importantly, the proof of Theorem 1.28 compares the first component of dG(u,χ(u)) with dF_u(ξ) and then absorbs the remaining terms into operators A and K. In the Rabinowitz operator the Lagrange multiplier τ is an independent variable, so dG contains no dχ term; hence the argument establishes Fredholmness of the frozen-χ operator, not of the true dF_u. Because Corollary 1.29 and the definition of CFH in Section 3 rely on the Fredholm property of dF_u, this is a load-bearing gap. To repair it, the author must either prove that the missing dχ term is a compact perturbation in the δ-weighted Sobolev spaces (with a concrete compactness argument), or modify the statement so that the frozen-χ operator is what is actually used in the definition of the moduli spaces.","section":"Lemma 1.21, Eq. (11); Theorem 1.28"},{"comment":"The proof of Lemma 2.7 uses the identity H(u(s,t)) = e^{r(s,t)} − 1 for arbitrarily large values of r, and the proof of Lemma 2.8 uses the fact that X_H and J are independent of r on the full cylindrical end [0,∞)×Σ. Lemma 1.25 only states H(r,x)=e^r−1 in an open neighborhood of Σ; it does not specify H on the rest of the cylindrical end. The gradient-type assumption ∇H=Λ introduced in Lemma 2.2 does imply H=e^r−1 globally on M+, but this implication is not stated or proved in the paper. Consequently, the radial-unboundedness argument, which is essential for the compactness theorem, currently relies on an unrecorded strengthening of the hypotheses. Please add an explicit statement (or proof) that under ∇H=Λ one has H=e^r−1 on [0,∞)×Σ, and similarly that X_H is r-independent there; alternatively, build the global form of H into Assumption A from the outset.","section":"Lemmas 2.7 and 2.8"},{"comment":"The proof of Lemma 2.1 introduces a constant a := −max{H(z) : z ∈ W\\{(r,x)∈U_ε : r>−r_0}} and asserts that W\\Σ = H^{-1}((−∞,0)), so that this maximum is strictly negative. This uses the sign convention that H is negative on the interior of the Liouville domain W. That sign condition is not part of Definition 1.24 of a defining Hamiltonian, nor is it stated explicitly in Lemma 1.25 or Assumption A. Since the uniform lower bound c>0 in Lemma 2.1 feeds directly into the χ-bound in Lemma 2.2 and hence into Theorem A and the compactness theory, the sign convention should be made an explicit hypothesis. This is a small but load-bearing missing assumption, and it should be stated before Lemma 2.1.","section":"Lemma 2.1"}],"minor_comments":[{"comment":"The right-hand side of equation (19) has a minus sign, whereas substituting ∇H = −JX_H into the flow equation ∂_s u + J(u)∂_t u + χ(u)∇H|_u = 0 gives ∂_s u + J(u)∂_t u = χ(u)J(u)X_H(u). The sign error does not affect the argument because the term is divided by α_n and tends to zero in the rescaling limit, but it should be corrected for consistency.","section":"Eq. (19)"},{"comment":"The statement 'da_H^γ(Λ|γ) = a_H(γ)' is slightly abusive because Λ|γ is not tangent to H, so da_H is not defined on it. The precise identity is da_γ(Λ|γ)=a(γ) for the unconstrained functional, combined with formula (4) for χ; please rephrase to avoid confusion.","section":"Proof of Lemma 2.2"},{"comment":"The symbol aH is used both for the constrained functional and for the action value a_H(γ); since the paper uses aH for the functional in Definition 1.9(i) and later for chain groups CF(aH,h), a typographically distinct notation for the functional (e.g., a^c) would improve readability.","section":"Notation"}],"recommendation":"major_revision","confidential_remarks":"The central construction is potentially valuable, and I see no circularity: the a priori bound is proved from first principles, and the Fredholm reduction targets a genuine external benchmark (RFH). The main obstacle is the treatment of dχ in the linearization; if that term is not compact, the Fredholm statement may need to be reformulated rather than patched. The global-H issue and the sign convention are straightforward fixes. I recommend major revision with the expectation that the author either supplies the missing compactness argument or adjusts the setup (e.g., by freezing χ or imposing a condition that makes dχ compact), and records the global cylindrical-end form of H as an explicit assumption."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe one-sentence take: this is a genuinely new construction — a Floer-type homology built from the symplectic area functional restricted to a mean-value zero hypersurface — but the Fredholm theory as written compares the wrong operator, so the main claims are conditional until that gap is fixed.\n\nWhat's actually new: the constrained functional a_H = a|_{h^{-1}(0)} and the non-local constraint factor χ are new to me, and the compactness bound under the gradient-type condition ∇H=Λ (Lemma 2.2) is a real analogue of the RFH multiplier bound. The idea of avoiding the Lagrange multiplier for the sake of a future product structure is sensible, and the paper is honest about the extra assumption and about what is deferred to a sequel.\n\nWhere it gets soft. Lemma 1.21's linearization formula omits the derivative of χ. As written, the displayed operator doesn't even map into u^*TH, since dχ(ξ)∇H has a normal component. The proof of Theorem 1.28 then identifies dF_u with the first block of dG, effectively freezing χ. The true operator differs by a rank-one term ξ ↦ dχ(ξ)∇H, which is likely compact in the weighted Sobolev spaces, so the Fredholm conclusion probably survives — but that step is not in the paper. That makes Theorem 1.28, and hence the finite-dimensionality of the moduli spaces, unproved as written.\n\nThe compactness section has a smaller but similar issue: Lemma 2.7 uses H=e^r−1 globally on the cylindrical end. This does follow from ∇H=Λ, since then dH = e^r dr and H=0 on Σ, but the paper never states that implication, and it's not among the lemma's hypotheses.\n\nThe rest — transversality by genericity, gluing, ∂²=0 — is deferred to \"usual arguments.\" That's normal for a preprint in this field, but it means the bona fide new analytic content is thinner than the theorems suggest.\n\nBottom line: the idea is good and probably correct, but the two gaps I named are load-bearing, not cosmetic. It deserves a serious referee, not a desk reject. I'd send it to review with a specific request to check the linearization and to state the global consequence of ∇H=Λ. I'd cite this paper if I worked on RFH or symplectic homology.","headline":"A genuinely new constrained symplectic area functional and a plausible Floer homology, but the Fredholm theory as written compares a frozen-χ operator, so the main theorems are conditional until that gap is fixed.","tokens_in":18169,"tokens_out":9368,"would_cite":true,"duration_ms":100655,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53D40","53D35"],"pacs":[],"model":"deepseek-v4-flash","headline":"Restricting the symplectic area functional to loops of zero Hamiltonian mean value defines a new Floer homology for Reeb orbits on Liouville boundaries, bypassing the Lagrange multiplier.","keywords":["constrained Floer homology","Rabinowitz Floer homology","Liouville domain","symplectic area functional","mean value constraint","Reeb orbits","Morse-Bott theory","compactness up to breaking"],"falsifier":"On the completion of a Liouville domain satisfying Assumption A but with a defining Hamiltonian modified in the compact interior so that $\\nabla H \\neq \\Lambda$, compute the constraint factor $\\chi(\\gamma) = -\\frac{da_\\gamma(\\nabla H|_\\gamma)}{dh_\\gamma(\\nabla H|_\\gamma)}$ along a family of loops $\\gamma_n \\in h^{-1}(0)$ whose images spend more and more time near a flat interior region. If $dh_{\\gamma_n}(\\nabla H|_{\\gamma_n}) \\to 0$ while $a(\\gamma_n)$ stays finite, then $|\\chi|$ is unbounded, the a priori bound (16) fails, and the whole compactness argument of Theorem 2.10 would collapse — showing the strengthened Weinstein condition is genuinely necessary.","tokens_in":17012,"feed_emoji":"🌀","tokens_out":12251,"duration_ms":129567,"temperature":0.7,"pith_summary":"This paper claims that a Floer homology exists for periodic Reeb orbits on the boundary of a Liouville domain, built from the symplectic area functional restricted to loops with vanishing Hamiltonian mean value. The result is a homology theory, called constrained Floer homology (CFH), whose chain groups coincide with Rabinowitz Floer homology's but which avoids the Lagrange multiplier and so is compatible with concatenation of loops. The paper proves the two prerequisites for a Floer theory: the gradient-flow moduli spaces are smooth finite-dimensional manifolds, and they are compact up to breaking. Both hold under an added geometric hypothesis — the Hamiltonian's gradient must equal the Liouville vector field — which controls the non-local constraint factor appearing in the gradient flow equation. If the construction works, it gives a route to a more intrinsic product structure on these homology groups.","feed_headline":"New Floer homology for Reeb orbits drops the Lagrange multiplier","feed_subtitle":"Shares Rabinowitz chain groups, avoids the multiplier, and may carry a product structure of its own.","key_machinery":"The load-bearing object is the constrained loop space $\\mathcal{H} = h^{-1}(0)$, a codimension-one Banach submanifold of the free loop space, with its tangent decomposition $T_\\gamma LM = T_\\gamma \\mathcal{H} \\oplus \\langle \\nabla H|_\\gamma \\rangle$. Along this submanifold the gradient of the constrained action is $\\nabla a_H(\\gamma) = J(\\gamma)\\partial_t \\gamma + \\chi(\\gamma)\\nabla H|_\\gamma$, where the constraint factor $\\chi(\\gamma) = -\\frac{da_\\gamma(\\nabla H|_\\gamma)}{dh_\\gamma(\\nabla H|_\\gamma)}$ is non-local: it depends on the whole loop. The Fredholm part of the proof arranges the linearized Rabinowitz operator as a direct sum of the linearized constrained operator and an index-zero model operator, so the constrained indices match. The compactness part rests on the strengthened Weinstein condition $\\nabla H = \\Lambda$, which yields the uniform bound $|\\chi(u(s))| \\leq C(x_\\pm, H)$ from Lemma 2.2; with it, a radially escaping sequence of cylinders is rescaled and shifted to produce a non-constant $J$-holomorphic curve whose $r$-component attains a local maximum, contradicting the maximum principle (Lemma 2.8).","core_discovery":"The paper's central claim is that the symplectic area functional restricted to the zero-level of the Hamiltonian mean value — the constrained function $a_H = a|_{h^{-1}(0)}$ — supports a Floer homology for the periodic Reeb orbits on the boundary $\\Sigma$ of a Liouville domain. Writing CFH for this constrained Floer homology, the paper proves that its gradient-flow moduli spaces $\\mathcal{M}(x_-, x_+)$ between critical points are smooth finite-dimensional manifolds (Theorem 1.28 and Corollary 1.29) and that they are compact up to breaking (Theorem 2.10), provided the Hamiltonian satisfies the gradient-type condition $\\nabla H = \\Lambda$. The construction shares the chain groups of Rabinowitz Floer homology, but because the constraint is imposed by restriction rather than by a Lagrange multiplier, the action functional stays additive under concatenation of loops — the feature that is meant to make an intrinsic product structure accessible. The paper's main technical achievements are the reduction of the Fredholm theory to Rabinowitz Floer homology and a bound on the non-local constraint factor $\\chi$ that arises from differentiating along the constraint.","pith_inferences":["The gradient-type hypothesis is a condition on the whole completion, not just on the collar: the standard defining Hamiltonian $H = e^r - 1$ automatically satisfies $\\nabla H = \\Lambda$ on the cylindrical end, so the hypothesis places constraints only on the extension of $H$ into the compact interior; whether such global $H$ exist for wide classes of Liouville domains is left open.","If the announced continuation paper establishes the product structure, the additivity of $a_H$ under concatenation would make CFH a better behaved carrier of the ring structure than RFH, where the multiplier is not additive under concatenation.","The proof of the $\\chi$-bound suggests that CFH might be definable under weaker assumptions as soon as a uniform lower bound on $dh_\\gamma(\\nabla H|_\\gamma)$ over $\\mathcal{H}$ is available; whether such a bound can hold without $\\nabla H = \\Lambda$ is a testable question that the paper does not settle."],"forward_implications":["CFH$(M, \\Sigma)$ is a well-defined homology theory for Liouville domains satisfying Assumption B: moduli spaces are finite-dimensional and compact up to breaking, so the Floer boundary map and its square-zero property follow from the standard machinery.","Because CFH shares its chain groups with Rabinowitz Floer homology, every CFH module carries the same underlying vector space as RFH; the difference is in the differential and in the absence of the Lagrange multiplier.","At critical points the constraint factor $\\chi$ coincides with the period functional (Theorem 1.12), so the generators of CFH are exactly the periodic Reeb orbits on $\\Sigma$, just as in RFH.","CFH is independent of the almost complex structure $J$ and the auxiliary Morse data, and is invariant along smooth families of defining Hamiltonians that stay within Assumption B (Theorem 3.4).","The Fredholm index of the constrained operator equals the RFH index, so the standard Morse–Bott treatment with an auxiliary Morse function and cascades applies without modification."],"supporting_citations":[{"why":"Supplies the Rabinowitz Floer homology framework — the Morse–Bott setup, the chain groups, and the Lagrange-multiplier bound (Prop. 3.2) that the $\\chi$-bound adapts to the constrained setting.","marker":"[CF09]"},{"why":"Provides the Banach-bundle and elliptic-regularity background (Chs. 8 and 12) and the derivative-bound Proposition 6.6.2 used to make the moduli spaces and the $L^\\infty$-bounds rigorous.","marker":"[AD14]"},{"why":"Gives the subharmonicity/maximum-principle computation (Lemma 4.1) that turns a radially escaping sequence of cylinders into a contradiction in Lemma 2.8.","marker":"[CFO10]"},{"why":"Supplies the $J$-holomorphic compactness and bubbling theorems (Theorem B.4.2) on which Lemma 2.5 and the compactness-up-to-breaking Theorem 2.10 rest.","marker":"[MS03]"},{"why":"Provides the index calculation for the model operator A (Prop. 2.16) that identifies the constrained Fredholm index with the Rabinowitz index in Theorem 1.28.","marker":"[Sch95]"}],"fun_headline_variants":["Constrained Floer homology for Reeb orbits, no multiplier","New Floer homology for Reeb orbits skips the multiplier","Reeb orbit Floer homology without the Lagrange multiplier","Symplectic area yields Floer homology, multiplier dropped","Floer homology for Reeb orbits, constrained and multiplier-free"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The construction requires the Hamiltonian to satisfy $\\nabla H = \\Lambda$, i.e., its gradient must equal the Liouville vector field, not merely be gradient-like, and this identity is what keeps the non-local constraint factor bounded; if it fails, the proof of compactness of the moduli spaces has no foundation.","fun_headline_variants_meta":{"raw":{"variants":["Constrained Floer homology for Reeb orbits, no multiplier","New Floer homology for Reeb orbits skips the multiplier","Reeb orbit Floer homology without the Lagrange multiplier","Symplectic area yields Floer homology, multiplier dropped","Floer homology for Reeb orbits, constrained and multiplier-free"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000152,"raw_usage":{"total_tokens":1207,"prompt_tokens":951,"completion_tokens":256,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":567,"completion_tokens_details":{"reasoning_tokens":167}},"tokens_in":567,"tokens_out":256,"duration_ms":3105,"temperature":1.0,"reasoning_tokens":167,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T19:13:51.791862+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On the completion of a Liouville domain satisfying Assumption A but with a defining Hamiltonian modified in the compact interior so that $\\nabla H \\neq \\Lambda$, compute the constraint factor $\\chi(\\gamma) = -\\frac{da_\\gamma(\\nabla H|_\\gamma)}{dh_\\gamma(\\nabla H|_\\gamma)}$ along a family of loops $\\gamma_n \\in h^{-1}(0)$ whose images spend more and more time near a flat interior region. If $dh_{\\gamma_n}(\\nabla H|_{\\gamma_n}) \\to 0$ while $a(\\gamma_n)$ stays finite, then $|\\chi|$ is unbounded, the a priori bound (16) fails, and the whole compactness argument of Theorem 2.10 would collapse — showing the strengthened Weinstein condition is genuinely necessary.","supporting_citations":[],"review_version":1}