Let f:M→Nf:M\to N be a between finite-dimensional without boundary, and let S⊆NS\subseteq N be an . The map ff is transverse to SS at x∈Mx\in M, written f⋔xSf\pitchfork_xS, if either f(x)∉Sf(x)\notin S, or

dfx(TxM)+Tf(x)S=Tf(x)N.df_x(T_xM)+T_{f(x)}S=T_{f(x)}N.

It is transverse to SS, written f⋔Sf\pitchfork S, if it is transverse at every x∈Mx\in M. Equivalently, ff and the inclusion S↪NS\hookrightarrow N are .

Normal-space formulation

At x∈f−1(S)x\in f^{-1}(S), compose the dfx:TxM→Tf(x)Ndf_x:T_xM\to T_{f(x)}N with the quotient map to Tf(x)N/Tf(x)ST_{f(x)}N/T_{f(x)}S. Transversality is equivalent to surjectivity of this composite. This says that ff supplies every direction normal to SS, even though dfxdf_x need not be surjective onto all of Tf(x)NT_{f(x)}N.

Preimage theorem

If f⋔Sf\pitchfork S, then f−1(S)f^{-1}(S) is an embedded submanifold of MM, with

Txf−1(S)=(dfx)−1(Tf(x)S)T_xf^{-1}(S)=(df_x)^{-1}(T_{f(x)}S)

and codimension equal to codim⁡NS\operatorname{codim}_N S. This is the transverse-preimage theorem.

Examples and non-examples

Every submersion is transverse to every embedded submanifold of its target. A constant map whose value lies in a positive-codimension submanifold is not transverse, because its differential contributes no normal directions. A map whose image misses SS is transverse by the vacuous clause.

Conventions and scope

Some authors state the condition only at points of f−1(S)f^{-1}(S); declaring it automatic elsewhere yields the same global notion. Boundary and corner versions require additional compatibility conditions not included here.

References
  1. M. W. Hirsch, Differential Topology, Springer, 1976. Springer DOI record. Relevant: Chapter 3, transversality and the inverse-image theorem.
  2. J. M. Lee, Introduction to Smooth Manifolds, 2nd ed., Springer, 2012. Springer DOI record. Relevant: Chapter 6, transversality.