-2
$\begingroup$

Given a smooth function $f:\mathbf R^n\to \mathbf R^m$ with $0$ as a regular value, I define the $(n-m)$ dimensional smooth manifold $M_f:=f^{-1}(0)$.

Let $f_0(x_1,...,x_n):=(x_1,...,x_m)$; $M_{f_0}$ is then simply $\mathbf R^{n-m}$. My question is: for every smooth $(n-m)$ dimensional manifold $\Sigma$, does there exist a function $g$ and a number $\delta>0$ such that $\forall \epsilon\in(0,\delta)$, $M_{f_0+\epsilon g}$ is diffeomorphic to $\Sigma$?

If this were the case then it could be said that all smooth manifolds are in the 'vicinity' of Euclidean space.

An example of this is: let $\Sigma=\mathbf S^2$ and $f_{\epsilon}(x,y,z):=\epsilon(x^2+y^2+z^2-1)+x$. $M_{f_0}=\mathbf R^2$, while $M_{f_\epsilon}$ has the topology of $\Sigma$ for all $\epsilon$ sufficiently small.

$\endgroup$
3
  • 4
    $\begingroup$ It is not true that $f^{-1}(0)$ is a smooth manifold unless $0$ is a regular value of $f$. Take any closed subset $C \subset \mathbb{R}^n$ and let $f:\mathbb{R}^n \to \mathbb{R}$ be the (smooth!) function sending $x$ to the squared Euclidean distance from $x$ to $C$. Then $C = f^{-1}(0)$ even though $C$ could be a horrible Cantor set. $\endgroup$ Commented Oct 1, 2022 at 21:43
  • 2
    $\begingroup$ @ViditNanda -- horrible :) ##### you have your way with words, nice! $\endgroup$
    – Wlod AA
    Commented Oct 2, 2022 at 4:22
  • 1
    $\begingroup$ @ViditNanda Edited. $\endgroup$
    – dennis
    Commented Oct 2, 2022 at 8:56

1 Answer 1

3
$\begingroup$

Under reasonable assumptions about $\Sigma$ the answer is yes. For example if $\Sigma$ is smooth and compact $(n-m)$-dimensional submanifold of $\mathbb{R}^n$ and it has trivial normal bundle*, that follows from the tubular neighborhood theorem. According to this theorem a neighborhood $U\subset\mathbb{R}^n$ of $\Sigma$ is diffeomorphic to $B^m\times\Sigma$. If $F:U\to B^m\times\Sigma$ is the diffeomorphism, $\pi:B^m\times\Sigma\to B^m$ is the projection onto the first factor, and $f=\pi\circ F$, then for every $y\in B^m$, $f^{-1}(y)$ is diffeomorphic to $\Sigma$.

This should also be true for smooth non-compact submanifolds with trivial normal bundle, because you can make tubular neighborhood gradually decrease as you move away. However, I did not think about that case carefully.


*There was a mistake in my previous answer and as observed by Andy Putman in his comment.

$\endgroup$
3
  • 1
    $\begingroup$ Aren't you also assuming that the normal bundle of $\Sigma$ is trivial? Otherwise, the tubular neighborhood won't split as a product. $\endgroup$ Commented Oct 1, 2022 at 20:54
  • 1
    $\begingroup$ @AndyPutman You are absolutely right. I wrote my answer without thinking enough. $\endgroup$ Commented Oct 1, 2022 at 21:10
  • $\begingroup$ @PiotrHajlasz Thanks but is this still true in the 'vicinity' of Euclidean space (as proposed in my question)? $\endgroup$
    – dennis
    Commented Dec 18, 2022 at 14:07

You must log in to answer this question.