I have arrived at an expression

$$\frac{f'(x+h)- \frac{f(x+h)-f(x)}{h}}{h}$$ for a compactly supported function $f \in C_C^{\infty}(\mathbb{R}).$

Now I was asking myself, whether we have uniform convergence to the second derivative of $f$ at point $x.$

Intuitively it is clear that there could be a second derivative involved, as the right summand in the nominator goes to $f'(x)$ and thus we have a difference quotient

$$\frac{f'(x+h)-f'(x)}{h}$$ which goes uniformly to zero, as $|f''|$ is bounded. Unfortunately, in this argument I am taking the limits separately which is not allowed I guess, so is there any way to make this argument rigorous?


Since $f$ is $C^{\infty}$ you have that, using Taylor expansion of $f$ and $f',$

$$f(x+h)=f(x)+f'(x)h+\frac12 f''(x)h^2+\frac16f(x)h^3\mathrm{lot},$$ and

$$f'(x+h)=f'(x)+f''(x)h+\frac12 f'''(x)h^2+\mathrm{lot}.$$ Thus,

$$\frac{f'(x+h)- \frac{f(x+h)-f(x)}{h}}{h}=\frac{f'(x)+f''(x)h-\frac{f'(x)h+\frac12f''(x)h^2}{h}}{h}+\mathrm{lot}$$

$$=\frac{f''(x)h-\frac12f''(x)h}{h}+\mathrm{lot}=\frac12 f''(x)+o(h).$$


It seems, surprisingly, that your expression converges to $f''(x)/2$ not $f''(x)$, as proven by @mfl. For example take $f(x)=x^2$ and compute the limit in $x=0$:

$$ \lim_{h \to 0}\frac{2h-\frac{h^2}{h}}{h}=2-1=1$$ but $f''(0)=2$.


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.