Show that the function is not differentiable at the point z

Join the discussion
Registration is free. Ask a follow-up in this thread, or start your own.
5 replies · 2K views
mathmari
Gold Member
MHB
Messages
4,984
Reaction score
7
Hi! Could you help me with the following?
Let [tex]g: R \rightarrow R[/tex] a bounded function. There is a point [tex]z \epsilon R[/tex] for which the function [tex]h: R[/tex] \ [tex]\{z\} \rightarrow R[/tex], where [tex]h(x)=\frac{g(x)-g(z)}{x-z}[/tex] is not bounded. Show that the function [tex]g[/tex] is not differentiable at the point [tex]z[/tex].

My idea is the following:
Let the function [tex]g[/tex] be differentiable at the point [tex]z[/tex], so [tex]lim_{x \rightarrow z}{ \frac{g(x)-g(z)}{x-z}}=L[/tex]. So there is a ε>0: [tex]|\frac{g(x)-g(z)}{x-z}-L|<ε[/tex].
[tex]|h(x)|=|h(x)-L+L| \leq |h(x)-L|+|L|=|\frac{g(x)-g(z)}{x-z}-L|+|L|<ε+|L|[/tex].
So the function [tex]h[/tex] is bounded. That cannot be true. So the function [tex]g[/tex] is not differentiable at the point [tex]z[/tex].

Is my idea right??
 
Physics news on Phys.org
mathmari said:
Hi! Could you help me with the following?
Let [tex]g: R \rightarrow R[/tex] a bounded function. There is a point [tex]z \epsilon R[/tex] for which the function [tex]h: R[/tex] \ [tex]\{z\} \rightarrow R[/tex], where [tex]h(x)=\frac{g(x)-g(z)}{x-z}[/tex] is not bounded. Show that the function [tex]g[/tex] is not differentiable at the point [tex]z[/tex].

My idea is the following:
Let the function [tex]g[/tex] be differentiable at the point [tex]z[/tex], so [tex]lim_{x \rightarrow z}{ \frac{g(x)-g(z)}{x-z}}=L[/tex]. So there is a ε>0: [tex]|\frac{g(x)-g(z)}{x-z}-L|<ε[/tex].
[tex]|h(x)|=|h(x)-L+L| \leq |h(x)-L|+|L|=|\frac{g(x)-g(z)}{x-z}-L|+|L|<ε+|L|[/tex].
So the function [tex]h[/tex] is bounded. That cannot be true. So the function [tex]g[/tex] is not differentiable at the point [tex]z[/tex].

Is my idea right??
Absolutely correct! (Sun)
 
Great! (Clapping)
 
Now I'm looking again the exercise..
There is a $ε>0$ such that $|\frac{g(x)-g(z)}{x-z}-L|<ε$ for $|x-z|<δ$, right? So $h$ is bounded for $x$ such that $|x-z|<δ$.
Do I have to show also that $h$ is bounded for $x$ such that $|x-z|>δ$, or is the above enough?
 
mathmari said:
Now I'm looking again the exercise..
There is a $ε>0$ such that $|\frac{g(x)-g(z)}{x-z}-L|<ε$ for $|x-z|<δ$, right? So $h$ is bounded for $x$ such that $|x-z|<δ$.
Do I have to show also that $h$ is bounded for $x$ such that $|x-z|>δ$, or is the above enough?
Yes, you need to show that if $h$ is unbounded, then it must be unbounded near $z$, not anywhere else. But if $|x-z|\geqslant \delta$ then $\left|\dfrac1{x-z}\right|$ is bounded by $1/\delta$. Also, you know that $g$ is bounded. So it follows that $h$ must indeed be bounded outside the region $|x-z|< \delta$.
 
Opalg said:
Yes, you need to show that if $h$ is unbounded, then it must be unbounded near $z$, not anywhere else. But if $|x-z|\geqslant \delta$ then $\left|\dfrac1{x-z}\right|$ is bounded by $1/\delta$. Also, you know that $g$ is bounded. So it follows that $h$ must indeed be bounded outside the region $|x-z|< \delta$.

So do I have to show both cases $ |x-z|< \delta$ and $ |x-z|\geq \delta$ ??