MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Consider the Hilbert space $L^2[0, 1]$ of square integrable functions on $[0, 1]$. Saying more carefully, there is one subtle detail: this space is defined as factor space by the functions which are have norm zero, in particular if we change value of $f(x)$ at one point, then in this factor space the two functions will be still the same. So formally speaking the value at points of elements from $L^2[0 1]$ is not defined.

Take point $x$ in $[0, 1]$. A continuous function $f$ can be evaluated at $x$: $ev_x (f) = f(x)$.

Question: Can the linear functional $ev_x$ be extended from continuous functions to $L^2[0, 1]$? In other words, does there exists a continuous linear functional $\Psi:L^2[0, 1]\to \mathbb{R}$, such that it coincides with $ev_x$ on the subspace of continuous functions?

Another way to Reformulation: let $i: C[0, 1] \to L^2[0, 1]$ be the inclusion and consider the dual map $i : (L^2[0,1])^{ * } \to (C[0,1])^{ * }$. The dual spaces involved are known: for $C[0, 1]$ it is space of functions of bounded variation, whereas $(L^2[0, 1])^{*} = L^2[0, 1]$. So we can ask whether $ev_x$ lies in the image of $i^{ * }$.

Motivation: in physics "coherent states" defined by this idea, (as far as I understand), e.g. approach by Rawnsley. But they are defined using "holomorphic polarization", where of course we can take value of holomorphic function at a point. So I am puzzled wether we really need holomorphic polarization.

share|cite|improve this question
Of course not: $ev_x$ is not continuous on $C([0,1])$ wrto the $L^2$ norm. There are continuous functions $f$ with arbitrarily small $L^2$ norm and such that $f(x)=1$. – Pietro Majer Feb 2 '12 at 12:52
Heh, you are right! Please write it as answer I will accept it. – Alexander Chervov Feb 2 '12 at 12:58
up vote 3 down vote accepted

We can't extend the evaluation at $x$ to a continuous linear functional on $L^2[0,1]$, because it is not continuous on $C([0,1])$ wrto the $L^2$ norm. For instance $f_n(t):=(1-nt)_+$ has $f_n(0)=1$ and $\|f_n\|_2\to 0$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.