gpt 6 astra pro attempt on Math-176

The Weinstein Conjecture

Definitions and mathematical statement

On a closed smooth $(2n+1)$-manifold $M$, a contact form is a one-form $\alpha$ with $\alpha\wedge(\,\mathrm d\alpha)^n$ nowhere zero. Its Reeb vector field $R_\alpha$ is uniquely specified by \[ \alpha(R_\alpha)=1,\qquad \iota_{R_\alpha}\,\mathrm d\alpha=0. \] A periodic orbit is a smooth map $\gamma:{\mathbb{R}}/T{\mathbb{Z}}\to M$, $T>0$, satisfying $\dot\gamma=R_\alpha(\gamma)$. The conjecture asserts the existence of such an orbit for every closed contact manifold and every contact form in all dimensions. Here closed means compact without boundary. No nondegeneracy, contractibility, or uniform bound on the period is imposed.

Formulation and concept references: [S1].

Short English statement

Must every Reeb flow on a compact contact manifold without boundary have at least one closed trajectory?

Research attempt

Current frontier.

The smooth closed-manifold assertion in the question is the one recorded in [S1], Section 5.1; results for singular contact forms in that paper must not be substituted for it. Taubes's theorem settles every contact form in dimension three. Hutchings gives both its precise statement and a survey of higher-dimensional special cases [E1], Theorem 1.3 and Sections 2--3. Those cases do not include every closed contact manifold in all dimensions. The investigation below does not assume nondegeneracy, fillability, a contractible orbit, or a bound on the period in the target.

Attack route.

A holomorphic-curve approach requires a nonvanishing invariant and compactness in the relevant contact geometry; a return-map approach needs a suitable section and a fixed-point mechanism. A third possibility is to approximate an arbitrary form by forms with explicit periodic fibers and pass to a limit. The decisive issue for that approach is a locally uniform upper bound on the selected periods. We prove the needed compactness lemma, construct controlled orbits for a restricted but infinite-dimensional class, and test whether either ingredient extends to arbitrary perturbations.

Attempt: exactly what approximation would need.

Fix an auxiliary Riemannian metric on the compact manifold. Suppose that contact forms $\alpha_j$ converge to a contact form $\alpha$ in $C^2$. Solving the pointwise nonsingular linear equations defining the Reeb field shows that $R_{\alpha_j}\to R_\alpha$ in $C^1$. Suppose further that there are periodic trajectories with initial points $p_j$ and periods $0<T_j\le C$, where $C$ is independent of $j$. Then $\alpha$ has a periodic trajectory.

Here are the two compactness details that this statement needs. Because $R_\alpha$ is nowhere zero and the manifold is compact, its norm has a positive minimum. In a finite collection of coordinate charts, the nearby fields have a common positive lower speed and common bounds on their values and first derivatives. The integral equation for their flows gives, uniformly for sufficiently small $t$, \[ \phi_j^t(p)-p=tR_{\alpha_j}(p)+O(t^2). \] Choose $t_0>0$ so that the remainder cannot cancel the first term and the flow remains in the chosen chart. Thus there are no periods in $(0,t_0)$, uniformly in large $j$. Passing to a subsequence gives $p_j\to p$ and $T_j\to T\in[t_0,C]$. Continuous dependence of flows on the vector field, uniform on bounded time intervals, implies \[ \phi_\alpha^T(p)=p. \] The resulting trajectory is nonconstant since its velocity never vanishes. This proves the lemma, including the lower-period issue.

Consequently a sufficient approximation statement is: for every $\alpha$ there is a $C^2$-convergent sequence of contact forms with periodic orbits whose periods are bounded by a constant depending on $\alpha$. Density of forms possessing an orbit, with no control of its period, does not establish this sufficient statement. The divergence $T_j\to\infty$ is a genuine failure of this compactness argument, even though the trajectories stay in a compact manifold.

Attempt: controlled periodic fibers under basic rescaling.

Consider a regular contact form $\alpha_0$ whose Reeb flow is a free circle action with period $T_0$. Write $\pi:M\to B$ for the quotient and $d\alpha_0=\pi^*\omega$, where $\omega$ is symplectic. This is the regular circle-bundle setting discussed in [E1], Section 2. Let $f$ be any smooth positive function on $B$, and put $\beta=(f\circ\pi)\alpha_0$. Suppress pullbacks in the following calculation. There is a unique horizontal vector field $Y$ such that \begin{equation}\label{wsref-0-65-71-31-36-39-2d-62-61-73-69-63} R_\beta=f^{-1}R_{\alpha_0}+Y, \qquad \iota_Yd\alpha_0\big|_{\ker\alpha_0}=f^{-2}df. \end{equation} Indeed the first formula gives $\beta(R_\beta)=1$. For a horizontal $V$, \[ d\beta(R_\beta,V)=-f^{-1}df(V)+f\,d\alpha_0(Y,V)=0. \] The remaining direction vanishes as well: evaluating the second equation on $Y$ gives $df(Y)=0$, while $df(R_{\alpha_0})=0$. Nondegeneracy on the horizontal distribution proves uniqueness and the Reeb equations.

At a critical point $b$ of $f$, the field $Y$ vanishes on the entire fiber. That fiber is therefore periodic, with period $T_0 f(b)$. Taking a minimum of $f$ gives the explicit bound \begin{equation}\label{wsref-0-65-71-31-36-39-2d-62-6f-75-6e-64} T\le T_0\min_B f. \end{equation} The calculation is not perturbative and requires no Morse hypothesis on $f$. It also checks the normalization under constant rescaling, for which Reeb time is multiplied by the same constant. Thus the desired compactness bound is available in this restricted class.

The attempt to extend the construction fails at a precise point. For a nonbasic function $h$ on $M$, a critical point of $h$ is not an entire critical circle; the horizontal part of $R_{h\alpha_0}$ need not vanish along its original fiber. Averaging $h$ along the circle action produces a basic function but is not an arbitrarily small approximation operation. The basic functions form a closed proper subspace in $C^2$: every such function satisfies $R_{\alpha_0}h=0$, and this equation survives limits. A function with a nonzero derivative in the fiber direction cannot be approximated in $C^2$ by basic ones. No density assertion follows from (eq169-basic).

Attempt: testing a global return section.

There is also an elementary obstruction to a tempting replacement argument. For $n\ge1$, a nonempty closed hypersurface $\Sigma^{2n}\subset M$ cannot be everywhere transverse to the Reeb field. If it were, $d\alpha|_\Sigma$ would be symplectic: the only kernel of $d\alpha$ in $TM$ is the Reeb line, and transversality identifies $T\Sigma$ with the symplectic quotient by that line. With its symplectic orientation, \[ 0<\int_\Sigma(d\alpha)^n =\int_\Sigma d\left(\alpha\wedge(d\alpha)^{n-1}\right)=0, \] a contradiction. Thus an everywhere-transverse compact return section without boundary is unavailable. Sections with boundary on periodic orbits are not ruled out, but assuming such a boundary would already assume the existence sought. This is an obstruction to that proof device, not to the Weinstein conjecture.

Stress test.

The dimension-one case is separate and immediate: each component of a closed one-manifold is a circle, and a nowhere-zero Reeb field traverses it in a finite positive time. The Stokes argument starts only in dimension three. The controlled-fiber construction needs a regular contact form to start with; neither existence of that form for a given contact structure nor approximation of arbitrary contact structures by regular ones has been proved here. Likewise, a general volume-preserving aperiodic flow is not a counterexample unless it satisfies the contact Reeb equations.

In the limit lemma the upper-period bound is an additional hypothesis, not a hidden conclusion of compactness of $M$. Allowing $\alpha_j$ to converge to a noncontact form would also invalidate the nonsingular linear-system argument. Every assumption used in that lemma is therefore kept explicit. The geometric computations are deductions within this attempt using a standard circle-bundle example; no claim of a newly solved higher-dimensional case or of literature priority is made.

Outcome.

Outcome: Exploratory approach with unresolved gap.

The attempt proves a bounded-period closure lemma, an exact period formula for positive basic rescalings of a regular contact form, and an obstruction to a closed global-section shortcut. These identify a possible approximation route and show why its easiest implementation does not reach arbitrary forms. A general construction of approximants with uniformly bounded periods, or a different existence mechanism covering the omitted contact geometries, remains unproved.

Sources

• S1 Eva Miranda and Cedric Oms, The singular Weinstein conjecture, Advances in Mathematics 389 $2021$, Section 5.1, Conjecture 5.1. https://arxiv.org/pdf/2005.09568. Catalogue formulation source.
• E1 Michael Hutchings, Taubes's proof of the Weinstein conjecture in dimension three, author survey, Theorem 1.3 and Sections 2--3. https://math.berkeley.edu/~hutching/pub/tw.pdf. Consulted 27 September 2026. Used for the three-dimensional theorem and regular contact circle-bundle examples, not an all-dimensional existence theorem.