---
title: Globally Tight Lagrangian Submanifolds
url: https://www.emergentmind.com/topics/globally-tight-lagrangian-submanifold
type: topic
---

# Globally Tight Lagrangian Submanifolds

A globally tight Lagrangian submanifold is, in the literature considered here, a Lagrangian submanifold whose transverse intersections with all relevant ambient translates attain the sharp lower bound given by the total \(\mathbb Z_2\)-Betti number. In homogeneous symplectic geometry this notion is formulated relative to a Lie group action, while in homogeneous Kähler geometry it is formulated relative to isometries. The central distinction is between **global tightness**, **local tightness**, and the more recent **infinitesimal tightness**; much of the modern literature establishes the latter two, whereas genuinely global results are rarer [1903.03717] [1308.3670].

## 1. Definition and intersection-theoretic meaning

Following Oh, the homogeneous-orbit framework defines a Lagrangian submanifold \(\mathcal L\subset M\) to be **globally tight** if for all \(g\in G\) such that \(\mathcal L\) and \(g(\mathcal L)\) intersect transversally,
\[
\#(\mathcal L\cap g(\mathcal L))=\mathrm{SB}(\mathcal L,\mathbb Z_2),
\]
where
\[
\mathrm{SB}(\mathcal L,\mathbb Z_2)=\sum_k b_k(\mathcal L;\mathbb Z_2).
\]
It is **locally tight** if the same equality holds for all \(g\) near the identity with transverse intersection [1903.03717]. In the irreducible compact homogeneous Kähler setting, the same formulas are used, but \(g\) ranges over isometries of the ambient manifold; in that literature, “tight” means local tightness unless otherwise specified [1308.3670].

The point of the definition is that tightness is the sharp case of the Arnold–Givental lower bound
\[
\#(\mathcal L\cap g(\mathcal L))\ge \mathrm{SB}(\mathcal L,\mathbb Z_2),
\]
when that bound applies [1903.03717]. Global tightness therefore encodes an exact intersection count under every allowed transverse translate, not merely a lower estimate.

This notion is intrinsically relative to an ambient symmetry class. In one strand of the literature the ambient symmetries are the elements of a Lie group acting by Hamiltonian symplectomorphisms; in another they are the isometries of a homogeneous Kähler manifold. The numerical condition is the same, but the quantifier “for all \(g\)” depends on the chosen category [1903.03717] [1308.3670].

## 2. Local and infinitesimal variants

A major refinement is the notion of **infinitesimal tightness**. Let \((M,\omega)\) carry a Hamiltonian \(G\)-action with \(\Ad^*\)-equivariant moment map \(\mu\), and let \(\widetilde X\) denote the fundamental vector field of \(X\in\mathfrak g\). An element \(X\in\mathfrak g\) is said to be transversal to \(\mathcal L\) if, whenever \(\widetilde X(x)\in T_x\mathcal L\), one has \(\widetilde X(x)=0\), and if
\[
f_{\mathcal L}(X)=\{x\in\mathcal L:0=\widetilde X(x)\in T_x\mathcal L\}
\]
is finite. Then \(\mathcal L\) is **infinitesimally tight** if for every transversal \(X\),
\[
\#f_{\mathcal L}(X)=\mathrm{SB}(\mathcal L,\mathbb Z_2).
\]
This replaces actual intersection counts by zero counts of infinitesimal generators [1903.03717].

The crucial equivalence is that if \(M\) is a homogeneous space with a \(G\)-invariant symplectic form, then
\[
\text{infinitesimally tight}\iff \text{locally tight}.
\]
The proof uses Weinstein’s neighborhood theorem: locally one identifies a neighborhood of \(\mathcal L\) with \(T^*\mathcal L\), compares the flow of \(\exp(tX)\) with the vertical component of \(\widetilde X\), and shows that for small \(t\),
\[
\#(\mathcal L\cap \exp(tX)(\mathcal L))=\#f_{\mathcal L}(X)
\]
[1903.03717].

What is not proved is equally important. The same paper explicitly does **not** claim that infinitesimal tightness implies global tightness, nor that local tightness implies global tightness. Thus infinitesimal tightness is a precise reformulation of the local problem, not a general solution of the global one [1903.03717].

## 3. Homogeneous orbit theory and explicit locally tight examples

The orbit-theoretic setting is organized by the moment map. If \(L\subset G\) is a Lie subgroup with Lie algebra \(\mathfrak l\), then for an orbit \(Lx\),
\[
T_x(Lx)=\{\widetilde X(x):X\in\mathfrak l\},
\]
and the isotropicity criterion is
\[
Lx \text{ isotropic } \iff \mu(x)\in (\mathfrak l')^\circ,
\]
where \(\mathfrak l'=[\mathfrak l,\mathfrak l]\). In the compact orthogonal setting this becomes
\[
Lx \text{ isotropic } \iff \mu(x)\perp \mathfrak l'.
\]
A Lagrangian orbit is an isotropic orbit of half dimension [1903.03717].

This criterion produces several homogeneous Lagrangian orbits. In complex flag manifolds, examples include an \(S^3\)-orbit in the full flag manifold \(\mathbb F(1,2)\), realized as a \(U(2)\)-orbit. In products of dual flags \(\mathbb F_\Theta\times \mathbb F_{\Theta^*}\), the diagonal orbit
\[
\mathcal L=\{(X,-X):X\in\Ad(U)(iH)\}
\]
is Lagrangian, and more generally so are the shifted graph orbits
\[
\Gamma\{-\Ad(m):\Ad(U)(iH)\to \Ad(U)(i\sigma(H))\}.
\]
For these families the paper proves infinitesimal tightness, hence local tightness [1903.03717].

The counting mechanism is Lie-theoretic rather than Floer-theoretic. In the flag examples, one studies the critical points of the height function
\[
\hat\mu(A)(x)=\langle A,x\rangle
\]
on an adjoint orbit. For regular \(A\), the singular points of the Hamiltonian vector field are counted by the Weyl-group quotient,
\[
\#\{\text{singularities of }\widetilde A\}=|\mathcal W/\mathcal W_\Theta|,
\]
and for a flag manifold this equals the sum of Betti numbers. That identity is what yields infinitesimal, hence local, tightness of the diagonal and shifted-diagonal Lagrangian orbits [1903.03717].

The paper also shows the limits of this method. In cotangent bundles of orthogonal Lie groups, for semisimple \(G\) the only isotropic left or right orbits are \(G(g,0)\), and their global or infinitesimal tightness is not studied. More broadly, the paper provides many locally tight homogeneous examples but no theorem upgrading them to global tightness [1903.03717].

## 4. Classification of tight Lagrangian homology spheres

A separate classification theorem treats compact tight Lagrangian submanifolds with the \(\mathbb Z_2\)-homology of a sphere in simply-connected irreducible compact homogeneous Kähler manifolds. Here “tight” means **locally tight**, not necessarily globally tight [1308.3670].

The theorem states that if \(L\) is a compact tight Lagrangian submanifold with the \(\mathbb Z_2\)-homology of a sphere, then \(L\) is an orbit of a compact subgroup, and the ambient manifold \(M\) and \(L\) are, up to biholomorphic homothety and congruence, exactly one of the following [1308.3670]:

| Ambient manifold \(M\) | Lagrangian \(L\) | Orbit description |
|---|---|---|
| \(Q_n=\mathrm{SO}(n+2)/\mathrm{SO}(2)\times \mathrm{SO}(n)\) | \(S^n\) | \(\mathrm{SO}(n+1)\)-orbit |
| \(\mathrm{SU}(n+1)/S(U(1)\times U(1)\times U(n-1))\) | \(S^{2n-1}\) | \(U(n)\)-orbit |
| \(\mathrm{SU}(3)/T^2\) | \(S^3\) | \(U(2)\)-orbit |
| \(\mathrm{Sp}(n+2)/U(2)\times \mathrm{Sp}(n)\) | \(S^{4n+3}\) | \((\mathrm{Sp}(1)\times \mathrm{Sp}(n+1))\)-orbit |
| \(F_4/T^1\cdot \mathrm{Spin}(7)\) | \(S^{15}\) | \(\mathrm{Spin}(9)\)-orbit |

Moreover, every such \(L\) coincides with a connected component of the fixed point set of an antiholomorphic isometric involution, so all classified examples are real forms in the generalized sense used there [1308.3670].

The proof reduces local tightness to Euclidean tautness. For \(M=G/H\), one constructs a \(K\)-equivariant full embedding
\[
\varphi:L\to V
\]
such that tightness of \(L\) implies that \(\varphi(L)\) is taut, and conversely, under the full-isometry-group hypothesis, tautness of \(\varphi(L)\) implies tightness of \(L\). Under the \(\mathbb Z_2\)-homology-sphere hypothesis, Chern–Lashof theory forces \(\varphi(L)\) to be a codimension-one round sphere, and an Ohnita–Gotoh-type inequality then reduces the classification to a short list of homogeneous pairs [1308.3670].

This theorem is structurally close to the global-tightness problem but does not solve it in general. The paper explicitly distinguishes globally tight from locally tight, and the classification theorem concerns the latter. It does note that real forms of Hermitian symmetric spaces were proved by Tanaka–Tasaki to be globally tight, but it does not assert that every example in its own classification is globally tight [1308.3670].

## 5. Rigidity notions adjacent to, but distinct from, global tightness

Several neighboring literatures study Lagrangian rigidity without proving global tightness. In the twistor-space setting of hyperbolic \(6\)-manifolds, Reznikov Lagrangians are shown to be monotone, to admit explicitly classified holomorphic discs, and to satisfy
\[
m_0=\pm 2\sqrt{q}[L_\Sigma],\qquad HF(L_\Sigma,L_\Sigma)=H_*(L;\mathbb C[q^{1/2}]).
\]
The same work explicitly states that it does **not** discuss tightness or global tightness; its conclusions are Floer-theoretic rigidity results, not intersection-theoretic tightness theorems [1106.3959].

In symplectic dynamics, invariant submanifolds can be forced to be Lagrangian or even graphs under dynamical hypotheses. For Tonelli Hamiltonian flows, one theorem states that a Lipschitz invariant \(d\)-dimensional submanifold with no conjugate points and suitable forward/backward equilipschitz recurrence is Lagrangian and is the graph of a \(C^1\) function, while another states that a closed Lagrangian covering in a cotangent bundle is actually a diffeomorphic graph. These are strong global rigidity statements, but they are not formulated as global tightness [1409.5204].

In strict nearly Kähler \(6\)-manifolds, every Lagrangian \(3\)-fold is orientable, minimal, and has vanishing Maslov \(1\)-form, yet the second variation formula is generally indefinite; if \(H^1(L^3,\mathbb R)\neq 0\), there are compactly supported variations decreasing volume. This shows that automatic minimality does not imply a local or global minimizing property of the kind one might associate with tightness [1408.6433]. A related extrinsic-rigidity result in the homogeneous nearly Kähler \(\mathbb S^6(1)\) proves
\[
\int_{M^3} \|h\|^2\left(\|h\|^2-\frac{5}{4}-\frac{3}{2}\Theta^2\right)\,dM \ge 0,
\]
with equality only for the totally geodesic \(\mathbb S^3(1)\) or the Dillen–Verstraelen–Vrancken Berger sphere; again, no global-tightness conclusion is drawn [1902.01641].

Analytic rigidity appears in other forms. Whitney spheres in \(\mathbb C^n\) and \(\mathbb{CP}^n\) are characterized by small-energy gap theorems under conditions such as \(\nabla^*T=0\) or \(\nabla^*\nabla^*T=0\), but these are energy-gap and classification results, not tightness theorems [2108.09657]. In holomorphic symplectic geometry, a bimeromorphically contractible compact holomorphic Lagrangian submanifold must be biholomorphic to \(\mathbf{CP}^n\), and a neighborhood of it is biholomorphically symplectomorphic to a neighborhood of the zero section in \(T^*\mathbf{CP}^n\); this is a neighborhood rigidity theorem rather than a global-tightness statement [2311.04360].

Constructive and background literatures also bear on the subject without addressing tightness directly. Smooth tropical hypersurfaces in \(\mathbb R^3\) can be lifted to smooth embedded Lagrangian submanifolds in \((\mathbb C^*)^3\), yielding many explicit examples but no tightness criteria [1804.01469]. The geometry of the Lagrangian Grassmannian supplies the natural target of the Lagrangian Gauss map, with
\[
T_L\Lambda(n)=S^2L^*,
\]
and furnishes a projective-incidence language for studying tangent-plane geometry, but it does not itself define global tightness [1805.04294]. Finally, the phrase “global Lagrangian” can refer, in the inverse problem of the calculus of variations, to a globally defined Lagrangian \(1\)-form for an ODE system; that is a different use of “Lagrangian” from the submanifold notion at issue here [1812.04270].

## 6. Conceptual status of the subject

The modern picture separates three levels. **Global tightness** is the strongest: the sharp Betti-number intersection count holds for every transverse ambient translate in the chosen symmetry class. **Local tightness** asks for the same equality only near the identity. **Infinitesimal tightness** reformulates the local count as a zero count for fundamental vector fields, and in homogeneous settings it is equivalent to local tightness [1903.03717].

What has been established most clearly is the local side. Homogeneous orbit theory provides many infinitesimally and locally tight Lagrangian orbits, and tight Lagrangian homology spheres in irreducible compact homogeneous Kähler manifolds have been classified under a strong topological hypothesis [1903.03717] [1308.3670]. By contrast, the transition from local or infinitesimal tightness to global tightness remains unproved in the cited works.

This suggests that global tightness is substantially stricter than several nearby rigidity phenomena that are sometimes conflated with it. Floer-theoretic nonvanishing, minimality, Hamiltonian graph rigidity, sharp extrinsic pinching, neighborhood normal forms, and energy-gap characterizations all single out distinguished Lagrangians, but none of them is equivalent in the present literature to the defining intersection formula
\[
\#(\mathcal L\cap g(\mathcal L))=\mathrm{SB}(\mathcal L,\mathbb Z_2)
\]
for all transverse translates. In that precise sense, the globally tight Lagrangian submanifold remains an intersection-theoretic notion whose strongest general results are still concentrated in highly symmetric settings [1903.03717] [1308.3670].

Source: https://www.emergentmind.com/topics/globally-tight-lagrangian-submanifold