I recognize the fact that the natural map from an infinite dimension vector space $V$ to it's double dual space $V^{**}$ need not necessarily to be surjective because we don't have that the $\dim V$ = $\dim V^{*}$.

However, what is a simple example to show that it is not surjective in this case?

My first instinct was to pick the vector space $V = (\mathbb{Z}/\mathbb{2Z})^{\mathbb{N}}$ just a simple vector space.

However, when picking up a functional in the dual space which wouldn't be isomorphic to the natural map for any $v \in V$ I wasn't able to write one down.

Can someone help me come up with an example / some intuition as to why this happens?



You can pick any infinite dimensional vector space. Say, take $V=k[x]$ to be just polynomials in one variable. It is countable-dimensional. If you take its dual, you get $V^*=k[[x]]$, which is already uncountable-dimensional. When taking double dual $V^{**}$ it will be even bigger. So there is no way $V$ will be isomorphic to $V^{**}$, you don't even have to worry about how exactly the map $V\to V^{**}$ looks like.

And this is true in general for infinite-dimensional vector spaces. Their linear duals are getting too big.

| cite | improve this answer | |
  • $\begingroup$ Is there any example which directly relates to what it looks like with the linear map as well? - And thanks so much, makes more sense now. $\endgroup$ – Nick R Mar 11 '14 at 16:14
  • $\begingroup$ In my example with polynomials, the canonical map $V\to V^{**}$ is a map, which sends each basis vector $x^n$ to the linear function $f_n$ on $V^*=k[[x]]$, defined by $f_n(\sum\limits_{i=0}^\infty c_ix^i)=c_n$. $\endgroup$ – Sasha Patotski Mar 11 '14 at 16:21

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.