5
$\begingroup$

trying to recall the $\epsilon, \delta$ definition of continuity, I came up with the following:

A function is continuous at $x$ if $\forall \epsilon > 0 \; \exists \; \delta > 0: |f(x-\delta) - f(x+\delta)| < \epsilon$.

This is very likely not equivalent to the Weierstrass' definition of continuity at $c$:

$\forall \epsilon > 0 \; \exists \; \delta > 0: |x-c| < \delta \Rightarrow |f(x) - f(c)| < \epsilon $.

Could you please point out where the first statement fails to be equivalent to Weierstrass'?

Many thanks!

$\endgroup$
  • 3
    $\begingroup$ First of all, in the first definition $f$ need not to be defined at $x$. $\endgroup$ – enzotib Jul 11 '14 at 11:08
  • $\begingroup$ A definition like this cannot fail per se as you are free to define any concept you like as long as it is self-consistent. However, it is not equivalent to the standard definition of continuous. $\endgroup$ – Winther Jul 11 '14 at 11:12
  • $\begingroup$ @Winther: thank you, edited. $\endgroup$ – vojta havlíček Jul 11 '14 at 11:18
6
$\begingroup$

The Weierstrass' definition considers all values of $x$ within distance $\delta$ of $c$.

Your definition only says that there are two equidistant points from $x$, of which the difference in function values is less than $\epsilon$.

Consider the function $f$ which is constantly 0 except for $f(0) = 1$. This is clearly not continous, but following your definition, for any $\epsilon > 0$ I can certainly pick any positive $\delta$ so that $$|f(-\delta) - f(\delta)| = 0 < \epsilon$$

$\endgroup$
4
$\begingroup$

Hint For both definition, consider the continuity in $0$ of

$$f(x) =\left\{ \begin{array}{ll} 132 & \text{if } x = 0 \\ 0 & \text{else} \end{array}\right. $$

$\endgroup$
3
$\begingroup$

This may seem natural to you, but it has some serious problems. For example, it says that every even function is continuous at $0$--that's not really acceptable, is it?

$\endgroup$
  • $\begingroup$ Original: Does it say so even for $f(x) = 1 (x > 0)$, $f(x) = $ for $(x \leq 0)$ though? Edit: Missed the "even". Thank you! $\endgroup$ – vojta havlíček Jul 11 '14 at 11:21
  • $\begingroup$ No, but that's not an even function. An even function satisfies $f(-x)=f(x)$, so you always have |$f(0-\delta)-f(0+\delta)|=0<\epsilon$ for every choice of $\delta$ and $\epsilon$ $\endgroup$ – MPW Jul 11 '14 at 11:25
  • $\begingroup$ Please see the edit above. Thank you! $\endgroup$ – vojta havlíček Jul 11 '14 at 11:26
  • $\begingroup$ I understand, thanks $\endgroup$ – MPW Jul 11 '14 at 11:27

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.