0
$\begingroup$

enter image description here

For non-analyticity, we can show that the Cauchy-Riemann equations don't hold anywhere. But here we can see that the Cauchy-Riemann equations hold when $y=2x$. I understand what the book did here but I just want to clarify something: If Cauchy-Riemann equations hold for $y=2x$, it doesn't necessarily mean that the function is differentiable. Do they just assume here that the function is differentiable along the line $y=2x$ and then talk about how the function won't be differentiable in any neighborhood of any complex number on the line $y=2x$ and so it won't be analytic anywhere? ( Since Cauchy-Riemann equations won't hold in the neighborhoods and hence not analytic). I just need to understand if they assumed that the function is differentiable along the line $y=2x$ or not. What's happening here? Thank you.

I cannot use "Sufficient Condition for Differentiability", since it is introduced later on in the book.

The problem below was solvable since I could use the "Sufficient Condition for Differentiability":

Show that the function $f(z) = x^2 - x + y + i(y^2 - 5y - x)$ is nowhere analytic but differentiable along $y = x + 2$. I tried doing this and so far the Cauchy-Riemann equations are only satisfied when we have $y = x + 2$. So when we have $y = x + 2$, all the demands for "Sufficient Condition for Differentiability" are met and so the function is differentiable along $y = x + 2$. However, the function won't be differentiable at every point in the neighborhood of any complex number along $ y = x + 2$, so it won't be analytic.

The question I have is, do we need differentiability along a curve to talk about differentiability at neighborhoods of points along the curve?

So far I've learnt that:

Analyticity at a point implies differentiability at a point but the converse is not true, Analyticity at every point in a domain D implies differentiability at every point in a domain D and the converse is true, Analytic at a point implies differentiable at a point implies Cauchy-Riemann equations hold at that point but the converses are not true.

Are all of these correct?

I know this is a lengthy question, but please help me understand this and clear my confusions. Thank you!

$\endgroup$

1 Answer 1

3
$\begingroup$

No, they did not assume that $f$ is differentiable at the pints of the form $x+2ix$. What happens is that the set $L=\{x+2ix\mid x\in\Bbb R\}$ is a line and therefore, for any $z\in L$, no disk centered at $z$ is contained in $L$. And so, for any disk centered at any point of $L$ there is a point of that disk that doesn't belong to $L$ and so, in particular, there is a point in that disk at which $f$ is not differentiable.

$\endgroup$
8
  • $\begingroup$ Yes, I understand that. We have points where f is not differentiable and hence f is not analytic anywhere. So we don't need differentiability along y = x + 2 to discuss differentiability at neighborhoods of points along y = x + 2? $\endgroup$ Dec 8, 2020 at 15:23
  • $\begingroup$ Yes, that is correct. $\endgroup$ Dec 8, 2020 at 15:24
  • $\begingroup$ Could you explain why? $\endgroup$ Dec 8, 2020 at 15:26
  • $\begingroup$ Isn't that what my answer is about? What more is needed? $\endgroup$ Dec 8, 2020 at 15:29
  • 1
    $\begingroup$ Yes, all of them are correct. $\endgroup$ Dec 8, 2020 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.