---
title: Submanifold-Genericity in Differential Geometry
url: https://www.emergentmind.com/topics/submanifold-genericity
type: topic
---

# Submanifold-Genericity in Differential Geometry

Submanifold-genericity is not a single uniform definition but a cluster of related notions in which genericity is formulated relative to submanifolds. In the cited literature, it appears as generic transversality of perturbations or translates of submanifolds, as a structural condition on real or CR submanifolds, as a Baire-category statement about ambient metrics that eliminate special submanifolds, and as a dynamical property of averages taken along dilates of submanifolds [1607.03220] [1207.3312] [1703.09240] [2507.15498]. This suggests a common organizing idea: the submanifold is either the geometric object on which genericity is tested, the locus that defines the relevant transversality problem, or the distinguished structure whose presence forces generic rigidity or singularity phenomena.

## 1. Genericity as transversality for maps on embedded submanifolds

A central formulation treats genericity through perturbations of a smooth map restricted to an embedded submanifold. In the setting of "Generic linear perturbations" [1607.03220], one fixes an \(n\)-dimensional smooth manifold \(N\), an open set \(U\subset \mathbb{R}^m\), an embedding \(f:N\to U\subset \mathbb{R}^m\), and a smooth map \(F:U\to \mathbb{R}^\ell\). For each linear map \(T\in L(\mathbb{R}^m,\mathbb{R}^\ell)\), one considers
\[
F_T\circ f:N\to \mathbb{R}^\ell,\qquad F_T:=F+T.
\]
The problem is to determine what happens for “generic” \(T\), meaning almost all linear perturbations.

The relevant transversality is formulated in multi-jet spaces. For a map \(g:N\to P\), its \(r\)-jet extension is \(j^rg:N\to J^r(N,P)\), and for configurations of \(s\) distinct points one uses
\[
N^{(s)}=\{(q_1,\dots,q_s)\in N^s\mid q_i\neq q_j\ \text{for }i\neq j\},
\]
together with the \(s\)-fold jet space \({}_sJ^r(N,P)\) and the multi-jet map \({}_sj^rg:N^{(s)}\to {}_sJ^r(N,P)\). A map \(g\) is transverse with respect to a submanifold \(W\subset {}_sJ^r(N,P)\) when \({}_sj^rg\) is transverse to \(W\).

The key class of targets is that of modular submanifolds. A submanifold \(W\subset {}_sJ^r(N,P)\) is modular when it is invariant under the natural \(\mathrm{Diff}N\times \mathrm{Diff}P\)-action, lies over a single coincidence stratum \(P_\pi\), and the tangent-space data \(E(g,q,W)\) form a \(J^r(N)_q\)-submodule in Mather’s algebraic description of jet-space tangent spaces [1607.03220]. The main statement is that almost all linear perturbations \(F_T\circ f\) are transverse with respect to a given modular submanifold. The submanifold aspect is essential: the source is not \(\mathbb{R}^m\) itself but the embedded submanifold \(f(N)\subset \mathbb{R}^m\).

This framework extends John Mather’s generic projections to a broader perturbative setting. It treats singularity types, self-intersections, and multi-point configurations as jet-transversality questions on an embedded submanifold, and it makes “genericity” mean transversality after varying ambient linear data rather than varying the submanifold intrinsically [1607.03220].

## 2. Group actions, Euclidean incidence geometry, and jet-space genericity

A second family of results studies genericity through motion of submanifolds by symmetries or through parameterized families of geometric probes. For a transitive Lie group action on a manifold \(M\), if \(A\) and \(B\) are embedded submanifolds of dimensions \(k\) and \(l\), then for a generic \(\sigma\in G\) the intersection \(\sigma(A)\cap B\) is transversal, hence a submanifold of dimension \(k+l-m\) or the empty set, where \(m=\dim M\) [1404.1760]. In this form, submanifold-genericity means that moving one submanifold through a sufficiently rich symmetry group puts it in general position relative to another.

An explicitly quantitative Euclidean version appears in "A Note on Generic Transversality of Euclidean Submanifolds" [1811.01160]. If \(\Sigma\subset \mathbb{R}^n\) is a \(d\)-dimensional \(C^1\)-embedded submanifold, the paper defines
\[
\mathscr{A}(\Sigma):=\bigg\{ a\in\mathbb{R}^n:\ {\rm volume}\,\Big\{ p\in\Sigma : \partial\mathbb{B}(a, |a-p|) \text{ is not transversal to }\Sigma \text{ at }p \Big\} > 0 \bigg\}.
\]
The theorem states that \(\mathscr{A}(\Sigma)\) is contained in a countable union of \((n-d-1)\)-dimensional affine planes [1811.01160]. Here the exceptional parameter set is not merely null or meagre; it is geometrically constrained by an explicit affine-rectifiable structure.

A more general jet-theoretic formulation is given in "A generalization of Thom's transversality theorem" [1001.2054]. If \(Y\subseteq J^r_{\mathrm{imm}}(D,M)\) and \(Z\subseteq J^r(D,N)\) are submanifolds with \(\sigma_Y\pitchfork \sigma_Z\), then the set of smooth maps \(f:M\to N\) for which
\[
f_*|_Y:Y\to J^r(D,N)
\]
is transverse to \(Z\) is residual in \(C^\infty(M,N)\). The same paper studies submanifolds \(A\subseteq J^s(M,N)\) defined by jet conditions and proves that, under the additional \(K\)-transversality hypothesis, for generic \(f\) the restriction \(g|_{(j^sf)^{-1}(A)}\) is also generic in the sense that
\[
j^r(g|_{f^*A})\pitchfork f^*A[B].
\]
This places submanifold-genericity inside a parametric Thom–Mather framework: the submanifold may itself be cut out as a jet-preimage, and genericity then propagates to restrictions along that submanifold [1001.2054].

## 3. Structural generic submanifolds in complex, CR, generalized complex, and Sasakian geometry

In several complex variables and CR geometry, “generic submanifold” is a structural term rather than a Baire-category one. A real \(p\)-plane \(\Pi\subset \mathbb{C}^n\) is generic when
\[
\Pi + J\Pi = \mathbb{C}^n,
\]
and a \(C^k\)-smooth real submanifold \(M\subset \mathbb{C}^n\) is generic when
\[
T_zM + JT_zM = \mathbb{C}^n,\qquad \forall z\in M.
\]
If \(\dim_{\mathbb{R}}M=n\), then \(T_zM\cap JT_zM=\{0\}\), and \(M\) is maximal totally real [1207.3312]. In "Subsets of full measure in a generic submanifold in \(\mathbb{C}^n\) are non-plurithin" [1207.3312], the main theorem states that if \(M\subset \mathbb{C}^n\) is a \(C^2\)-smooth generic submanifold and \(I\subset M\) has measure zero in \(M\), then \(M\setminus I\) is non-plurithin at any point of \(M\). The proof uses attached analytic discs and the geometry of generic tangent spaces.

A related but distinct CR-geometric use occurs for Sasakian manifolds. In "A Frankel type theorem for generic submanifolds of Sasakian manifolds" [2006.07922], a submanifold \(N\subset M\) is called generic when \(\xi\) is nowhere normal to \(N\) and
\[
\phi(TN^\perp)\subset TN.
\]
This is weaker than the customary Sasakian condition requiring \(N\) to be tangent to the Reeb field. For such a generic submanifold of codimension \(p\), the induced pair \((HN,J)\) defines a CR structure on \(N\) of CR codimension \(p+1\), and the paper proves Frankel-type intersection theorems under lower bounds on the indices of the characteristic Levi forms determined by normal directions [2006.07922].

Generalized complex geometry contributes a different structural variant. "A note on submanifolds and mappings in generalized complex geometry" shows that one can characterize when a linear subspace or submanifold has an induced generalized complex structure, give a smoothness criterion for the induced structure, dualize the results to submersions, and identify generalized Kähler submanifolds as precisely the common invariant submanifolds of the two classical complex structures of the generalized Kähler manifold [1412.1217]. In this sense, submanifold-genericity is tied to compatibility with ambient complex, Poisson, or CR data rather than to residual subsets of an ambient parameter space.

## 4. Generic ambient metrics and the disappearance of totally geodesic submanifolds

A major Baire-category use of submanifold-genericity concerns the nonexistence of special submanifolds for typical ambient metrics. For a compact smooth manifold \(M\) of dimension at least \(4\), "Random Manifolds have no Totally Geodesic Submanifolds" proves that for any finite \(q\ge 2\), the set of \(C^q\) Riemannian metrics on \(M\) with no nontrivial immersed totally geodesic submanifolds contains an open dense subset [1703.09240]. The paper strengthens this by introducing partially geodesic \(l\)-planes and proving that, for each \(l\in\{2,\dots,n-1\}\), the set of metrics with no partially geodesic \(l\)-planes is open and dense [1703.09240]. Here the submanifold is absent generically because the tangent-plane conditions necessary for total geodesy fail after perturbation.

The dimension-three case is treated separately in "Random 3-Manifolds Have No Totally Geodesic Submanifolds" [2404.01581]. Earlier work had produced only a dense \(G_\delta\) subset of metrics without immersed totally geodesic surfaces. The new result shows that for a compact smooth \(3\)-manifold and any finite \(q\ge 3\), the set of metrics with no immersed totally geodesic surfaces contains a set that is open and dense in the \(C^q\)-topology [2404.01581]. The argument uses the generic plane operator
\[
G(P):=\max_{\substack{v,w\in P\\ |v|=|w|=1}}
\Big\{\big|R_v(w)^{P^\perp}\big|,\ \big|(\nabla_vR_v)(w)^{P^\perp}\big|\Big\},
\]
defines \(G\)-generic metrics by the condition \(G(P)>0\) for all tangent \(2\)-planes \(P\), and proves that the set of \(G\)-generic metrics is open and dense [2404.01581].

In both dimensions \(n\ge 4\) and \(n=3\), the phrase “random manifold” is used in the Baire-category sense, not through a probability measure on the space of metrics. The common conclusion is that totally geodesic submanifolds are nongeneric ambient phenomena: they persist in symmetric or rigid geometries, but an open dense set of metrics removes them [1703.09240] [2404.01581].

## 5. Minimal, prescribed-mean-curvature, and area-minimizing submanifolds under generic metrics

Submanifold-genericity also appears in variational geometry through transversality properties of minimal and prescribed-mean-curvature immersions. In "Generic Transversality of Minimal Submanifolds and Generic Regularity of Two-Dimensional Area-Minimizing Integral Currents" [1901.05148], for a smooth manifold \(N\) with a fixed smooth metric \(g_0\) and a smooth submanifold \(\Gamma\subset N\), a generic smooth metric \(g\) conformal to \(g_0\) has the property that every simple \(g\)-minimal immersion of a closed manifold into \(N\) is transverse to \(\Gamma\) and self-transverse. The paper strengthens both conclusions to strongly transverse and strongly self-transverse versions, defined by multi-point transversality of
\[
F^{(k)}:\widetilde{\Delta}^kM\to N^k
\]
to \(\Delta^k\Gamma\) and \(\Delta^kN\), respectively [1901.05148].

The same work shows that the transversality theorem extends from minimal immersions to hypersurfaces of constant mean curvature and, more generally, to hypersurfaces of prescribed mean curvature [1901.05148]. The analytic mechanism combines bumpy metric theory, Runge-type approximation for elliptic operators, and parametric transversality in spaces of conformal factors.

A further consequence concerns geometric measure theory. For a generic ambient metric, every \(2\)-dimensional locally area-minimizing integral cycle has support equal to a smoothly embedded minimal surface, and the same holds for \(2\)-dimensional area-minimizing flat chains mod \(2\), where the support is a smooth embedded minimal surface with multiplicity \(1\) [1901.05148]. In this setting, genericity acts on the ambient metric but regularizes the submanifold-like supports of minimizers.

A different variational enlargement is developed in "Frame bundle approach to generalized minimal submanifolds" [1601.02248]. There the codimension-\(q\) family of shape operators associated with all orthonormal normal frames is encoded through generalized symmetric functions
\[
\chi_A(t)=\det(I+tA)=\sum_{|u|\le n}\sigma_u(A)\, t^u
\]
and generalized Newton transformations \(T_u\). The resulting functional
\[
\mathcal F_u(M)=\int_M \sigma_u\, d\mu
\]
leads to \(u\)-minimality, with Euler–Lagrange equation \(R_u+W_u=S_u\), and in a space form of sectional curvature \(c\) this reduces to
\[
c(n+1-|u|)H_u=S_u
\]
[1601.02248]. This does not formulate Baire-genericity, but it enlarges the class of extrinsic curvature conditions that can play the role of “generic” higher-order mean curvature equations in arbitrary codimension.

## 6. Rigidity and singularity theories centered on distinguished submanifolds

Several papers use submanifolds as the carriers of rigidity or singularity statements that are then promoted to generic conclusions. In "Singular genuine rigidity" [1803.06395], the notion of genuine rigidity is extended by allowing mild singularities in the higher-dimensional extensions of isometric immersions. A principal consequence is that any compact \(n\)-dimensional submanifold of \(\mathbb{R}^{n+p}\) is singularly genuinely rigid in \(\mathbb{R}^{n+q}\) for
\[
q < \min\{5,n\} - p.
\]
The paper emphasizes that the singular theory is simpler and more natural than the regular one, while removing the technical codimension assumptions needed in the regular case [1803.06395].

In sub-Riemannian geometry, "On Weyl's type theorems and genericity of projective rigidity in sub-Riemannian Geometry" studies distributions \(D\subset TM\) as the geometric substrate. It proves that the Weyl projective rigidity analogue holds in the real-analytic category for all sub-Riemannian metrics on distributions whose complex abnormal extremals have minimal order, and in the smooth category under corresponding hypotheses on nilpotent approximations [2001.08584]. The same paper states that, in the real-analytic category, distributions for which all sub-Riemannian metrics are Weyl projectively rigid are generic, and that Weyl projectively rigid sub-Riemannian metrics on a given bracket generating distribution are also generic [2001.08584]. Here the relevant “submanifold” is the distribution itself, viewed as a subbundle whose abnormal geometry controls rigidity.

Lorentzian singularity theory offers another use. "Genericity of singularities in spacetimes with weakly trapped submanifolds" proves that, within the class of stably causal spacetimes of dimension \(n\ge 3\) satisfying the timelike convergence condition and containing a codimension-two spacelike weakly trapped closed submanifold, the existence of causal incomplete geodesics is a \(C^\infty\)-generic feature; an analogous statement holds for weakly trapped closed spacelike submanifolds of any codimension \(k>2\) under a modified curvature condition [2309.03421]. The follow-up "On the genericity of singularities in spacetimes with weakly trapped submanifolds" sharpens the picture: in strong Whitney topologies, singular Lorentzian metrics around a fiducial metric possessing a weakly trapped submanifold \(\Sigma\) are not really generic but are nevertheless prevalent in a sense defined there, while for initial data sets containing MOTS the paper obtains true genericity of null geodesic incompleteness around suitable initial data sets [2406.09651].

These rigidity and singularity results suggest a recurrent pattern. A distinguished submanifold or subbundle—an immersed Euclidean submanifold, a bracket-generating distribution, a weakly trapped surface, or a MOTS—encodes the geometry strongly enough that nearby ambient structures generically lose flexibility or develop incompleteness [1803.06395] [2001.08584] [2309.03421] [2406.09651].

## 7. Ergodic averages along submanifolds and the limits of \(L^\infty\)-genericity

A recent dynamical use makes the term itself explicit. "Higher-Dimensional Moving Averages and Submanifold Genericity" defines, for a measure-preserving \(\mathbb{R}^d\)-action on \((X,\mu)\), a compact \(m\)-dimensional \(\mathscr C^1\)-submanifold \(M\subset \mathbb{R}^d\), and a class \(F\subset L(X)\), the notion that a measure \(\nu\) is \((M,F)\)-generic if for \(\nu\)-almost every \(x\) and every \(f\in F\),
\[
\frac{1}{t^m \operatorname{vol}_m(M)}\int_{tM} f(a.x)\, d\operatorname{vol}_m(a)\to \mu(f)\qquad\text{as }t\to\infty.
\]
It is \((m,F)\)-generic if this holds for every compact \(m\)-dimensional \(\mathscr C^1\)-submanifold \(M\subset \mathbb{R}^d\) [2507.15498].

The same paper generalizes one-dimensional results of Bellow, Jones, and Rosenblatt to \(\mathbb{Z}^d\)- and \(\mathbb{R}^d\)-actions by giving necessary and sufficient coordinatewise cone conditions \((C_i)\) and \((\tilde C_i)\) for maximal inequalities and pointwise convergence of moving averages over families of boxes [2507.15498]. The application to submanifolds is negative for bounded functions. If a compact \(m\)-dimensional \(\mathscr C^1\)-submanifold \(M\subset \mathbb{R}^d\) contains a non-empty open subset lying in an \(m\)-dimensional affine subspace \(\pi\) that does not contain the origin, then for an ergodic and aperiodic measure-preserving \(\mathbb{R}^d\)-action the invariant measure \(\mu\) is not \((M,L^\infty(X))\)-generic [2507.15498]. Consequently, no ergodic and aperiodic action is \((m,L^\infty(X))\)-generic for any \(m=1,\dots,d-1\) [2507.15498].

This identifies a precise obstruction: local flatness of the averaging submanifold can force failure of pointwise convergence for bounded measurable functions. In this usage, submanifold-genericity is neither a transversality condition nor a structural property of tangent spaces, but a pointwise ergodic theorem along dilates of submanifolds. The contrast with the positive \(C_c^\infty\)-genericity results cited there underscores that the admissible function class is part of the notion itself [2507.15498].

Source: https://www.emergentmind.com/topics/submanifold-genericity