I am trying to show

Consider $F\colon C[a,b] \rightarrow R$ ; $F(f)=f(t_0)$ where $a<t_0< b$ . Show that this linear functional is not continuous under $\|\cdot\|_1$ on C[a,b]

I have tried a proof that is apparently wrong and I would like help finding out where the flaw is. I try to show the functional is discontinuous by constructing a sequence of functions such that $f_n \rightarrow f$ under $ \|.\|_1$ but $F(f_n)\not\to F(f)$ where $f_n$ and $f$ are continuous on [a,b].

$f_n(x)= n(x-t_o)+1, \, x\in [-1/n +t_0,t_0]$
$f_n(x)= n(-x+t_o)+1, \, x\in [t_0,t_0 +1/n]$

$f_n(x)= 0$, otherwise

Then, $\|f_n-0\|_1=\|f_n\|_1=(2/n)\times 1/2=1/n \to 0$ as $n\to \infty$ So $f_n \to f$ under $\|.\|_1$, where $f=0$ on $[a,b]$.

However $F(f)=f(t_0)=0$ and $F(f_n)=f_n(t_0)=1 $ for all $n$. Thus $|F(f_n)-F(f)|=1$. Therefore $f_n \rightarrow f$ under $ \|.\|_1$ but $F(f_n)\not\to F(f)$.

  • $\begingroup$ Why do you think your proof is wrong? Perhaps you should first prove $\;f_n\in C[a,b]\;$ ?? BTW, is that norm of yours the integral norm $\;\int_a^b |f(x)|dx\;$ ? $\endgroup$ – DonAntonio Mar 21 '14 at 15:55
  • $\begingroup$ I have readed it more than five times, and I can't see any major flaws. Perhaps you should define $f_n=0$ if $[t_0-1/n,t_0+1/n]\not\subset[a,b]$. $\endgroup$ – ajotatxe Mar 21 '14 at 16:09

Your proof is correct, and it is the natural idea to attack this problm. One could criticize that you don't explicitly show that the $f_n$ are continuous, but that's a minor point that is saved with a sentence noting that the only points where there could be any doubt are $a,t_0, b$ and that in the three cases the side limits agree.


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.