0
$\begingroup$

I'm reading a theorem in Munkres' "Analysis on Manifolds" that says

Let $A$ be an open subset of $\mathbb{R}^n$, let $f:A\to \mathbb{R}^n$ be of class $C^1$. If $Df(a)$ is non singular, then there exists $\alpha>0$ such that the inequality \begin{equation}|f(x_0)-f(x_1)|\geq \alpha |x_0- x_1|\end{equation}

holds for all $x_0,x_1$ in an open cube $C(a;\epsilon)$ centered at $a$. It follows that $f$ is one-to-one on this open cube.

WHY does it follow from the conclusion that $f$ is one-to-one? Very basic question, but I don't see it. Thanks!

$\endgroup$
  • 1
    $\begingroup$ If $x_0 \ne x_1$, what do you know about $\alpha|x_0 - x_1|$? $\endgroup$ – Anthony Carapetis Jun 12 '17 at 12:11
1
$\begingroup$

A function being one-to-one implies that it never maps distinct elements of its domain to the same element of its codomain.

Let's assume that this function is not one-to-one. Then there exists at least one pair of points, $y_0, y_1$ s.t. $f(y_0)=f(y_1)$. Namely, this implies $f(y _0)-f(y_1)=0$, and $y_0-y_1 \ne 0$. Plugging this into the relation $$|f(y_0)-f(y_1)| = 0 \ge \alpha |y_0-y_1| > 0$$ Which is a contradiction.

$\endgroup$
2
$\begingroup$

If $f(x_0)=f(x_1)$, then $0\leqslant\alpha|x_0-x_1|\leqslant\bigl|f(x_0)-f(x_1)\bigr|=0$, and so $|x_0-x_1|=0$. Of course, this means that $x_0=x_1$.

$\endgroup$

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.