I know that this is an isomorphism. I can see why the function is injective, but not why it is surjective.


1 Answer 1


Note that $\dim(V^{**}) = \dim(V)$. By the rank-nullity theorem, any injective map between vector spaces of the same dimension is necessarily surjective.

  • $\begingroup$ For a given $\psi \in V^{**}$, how can one determine a vector $v \in V$ with $\Phi(v)=\psi$? $\endgroup$ Mar 7, 2017 at 2:42
  • $\begingroup$ Here's one approach: let $e_1,\dots,e_n$ denote a basis of $V$. Let $\alpha_1,\dots, \alpha_n$ denote the associated dual basis of $V^*$. Then $$ v = \sum_{i=1}^n \psi(\alpha_i)e_i $$ Defining an explicit inverse will ultimately depend (in some way) on the fact that we're considering a finite dimensional space. $\endgroup$ Mar 7, 2017 at 2:45
  • $\begingroup$ Thanks! For beginners like me, here is how to derive this (could you please point out if the following has errors?): We're looking for $v \in V$ with $\psi(f)=f(v)$ for every $f \in V^{*}$. For $f=\sum_{i=1}^{n}\lambda_i.\alpha_i$ and $v=\sum_{j=1}^{n}\mu_j.e_j$ we have: $(\sum_{i=1}^{n}\lambda_i.\alpha_i)(\sum_{j=1}^{n}\mu_j.e_j)=\sum_{i=1}^{n}\lambda_i.\psi(\alpha_i)$, therefore $\sum_{i=1}^{n}\lambda_i.\mu_i = \sum_{i=1}^{n}\lambda_i.\psi_(\alpha_i)$ . Since this is satisfied for $\mu_i = \psi(\alpha_i)$, the unique $v$ we are looking for is $v=\sum_{i=1}^{n}\psi(\alpha_i).e_i$. $\endgroup$ Mar 7, 2017 at 16:55

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.