6
$\begingroup$

Let $f\colon M \to N$ and $g\colon A \to N$ be two smooth maps between manifolds $A,M,N$.

Can one perturb $f$ to be transverse to $g$ (without touching $g$)?

Transverse meaning: For every $y\in f(M)\cap g(A)$ and every $x\in f^{-1}(y)$ and $a\in g^{-1}(y)$, we have $$ Df(T_x M) + Dg(T_a A) = T_yN. $$

I always assumed that this was true, but searching for references, I found exercise 14 in Section 3.2 of the book by Hirsch, "Differential Topology" stating: "Is it true (as seems likely) that the set $\{f\in C^\infty(M;N): \text{$f\cap g$ transversely}\}$ is residual in $C^\infty(M;N)$ and open if $g$ is proper?"

The exercise is marked with three stars meaning "Three-star 'exercises' are problems to which I do not know the answer."

$\endgroup$
  • $\begingroup$ Don't know if the general situation can be solved, but in some particular cases, one could perturb f(M) to be a submanifold of N and suppose its normal bundle is trivial, then composing g with the projection to the fiber and using Sard's theorem, we would get that most sections in this normal bundle are transverse to g. $\endgroup$ – Klaus Niederkrüger Jan 13 '19 at 12:29
  • $\begingroup$ May be you should add your comment in the question... This is interesting question.... What does it mean to say a set is residual in $C^{\infty}(M,N)$?? $\endgroup$ – Praphulla Koushik Jan 13 '19 at 13:24
5
$\begingroup$

I'm not sure what exactly you mean by perturb, but you can always make $f$ transverse to $g$ by a homotopy and this is often enough. This is proved in Section IV.2 of Kosinski's "Differential Manifolds" book, in particular see Corollary IV.2.5.

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ By the multijet transversality theorem... $\endgroup$ – John Klein Jan 13 '19 at 15:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.