1

A First Course in Complex Analysis by Matthias Beck, Gerald Marchesi, Dennis Pixton, and Lucas Sabalka Exer 2.18

enter image description here



  1. Is my understanding of the following solution right, and why/why not?

At a Winter 2017 course in Oregon State University, there's a solution given: It seems to be saying that $f$ is not continuous and therefore conditions of Cauchy-Riemann do not hold because of some relationship between the continuity of $f$ and then parts of $f_x$ and $f_y$ and somehow the assumptions of C-R are equivalent to continuous differentiability.

Note: I assume the 'Theorem 2.15(b)' refers to an old edition, w/c is not relevantly different from the current edition, in w/c Cauchy-Riemann is in Thm 2.13.

  1. Here is my solution. Where have I gone wrong, if anywhere, and why?

Rewrite $f(z) = f(x,y) = 1_{xy \ne 0}(x,y)$ s.t.

$$f_x(0,0) := \lim_{\Delta x \to 0} \frac{1_{00+\Delta x 0 \ne 0} - 1_{00 \ne 0}}{\Delta x} = \lim_{\Delta x \to 0} \frac{0-0}{\Delta x} = \lim_{\Delta x \to 0} 0 = 0$$

Similarly, $f_y(0,0) = 0$.

$$\therefore, f_x(0,0) = 0 = -i0 = -if_y(0,0)$$

Now $f_x(0,0)$ may be defined, but $f_x$ neither is continuous at $0$ nor exists on any disc centred at $0$ because $f_x$ does not exist along $x_0=0$:

$$f_x(0,y_0) = \lim_{\Delta x \to 0} \frac{1_{0y_0+\Delta x y_0 \ne 0} - 1_{0y_0 \ne 0}}{\Delta x} = \lim_{\Delta x \to 0} \frac{1_{\Delta x y_0 \ne 0}-0}{\Delta x} = \lim_{\Delta x \to 0} \frac{1}{\Delta x} \ \text{, w/c dne.}$$

Similar holds for $f_y$ along $y_0 =0$.

$\therefore,$ Cauchy-Riemann holds at $z=0$ but $f$ is not differentiable at $z=0$ because $f_x$ is not continuous at $z=0$ because $f_x$ does not exist along $x_0=0$.

BCLC
  • 13,459
  • 2
    Partial derivative $f_x(z)$ does not exist at points $z=it$, $t$ small but non-zero. Neither does $f_y(z)$ when $z$ is a small non-zero real number. Therefore there is no open disk around the origin where those partial derivative would exist. – Jyrki Lahtonen Aug 02 '18 at 04:36
  • 1
    Why would you assume you have done something wrong?? – Eric Wofsey Aug 02 '18 at 04:39
  • @JyrkiLahtonen thanks! And that's what I did right? – BCLC Aug 02 '18 at 04:53
  • @EricWofsey edited. thanks – BCLC Aug 02 '18 at 12:10
  • 1
    Recall that being complex differentiable means the Cauchy--Riemann equations holds and that the function is real differentiable at that point. This function is not real differentiable at the origin. – Pedro Aug 08 '18 at 05:26
  • @PedroTamaroff Thanks! Post as answer? ^-^ – BCLC Aug 08 '18 at 06:14
  • @PedroTamaroff Oh wait, you're stating something stronger than Cauchy-Riemann Thm 2.13(a)? From 2.13(a), if the $f(x,y)$'s partial derivatives didn't exist, then $f(z)$ isn't differentiable. But now you're saying that if $f(x,y)$ isn't differentiable then $f(z)$ isn't differentiable? As I recall, if $f(x,y)$ isn't differentiable, then $f(x,y)$'s partial derivatives exist and the converse is not true, so to change the conclusion in Thm 2.13(a) from '$f(x,y)$'s partial derivatives exist and C-R' to '$f(x,y)$ isn't differentiable and C-R' is to strengthen? – BCLC Aug 08 '18 at 06:21

2 Answers2

2

That a function $f$ is complex differentiable at a point $z_0$ means that the derivative $$ f'(z_0) = \lim_{\mathbb{C} \ni h \to 0} \frac{f(z+h)-f(z)}{h} $$ exists.

Now, if $f$ (i.e. the real and imaginary parts of $f$) is real differentiable, and Cauchy-Riemann's equations are satisfied at $z_0$, this implies that $f$ is complex differentiable at $z_0$.

Note also, that existence of partial derivatives is not enough to conclude real differentiability!

mrf
  • 43,639
1

Recall that being complex differentiable means the Cauchy--Riemann equations holds and that the function is real differentiable at that point. This function is not real differentiable at the origin.

With respect to the second comment: if the partial derivatives exist in a neighborhood of the point and are continuous, then the function is real differentiable, so this plus Cauchy--Riemann gives complex differentiablity, in light of the previous paragraph.

Pedro
  • 122,002