There exist the famous theorem about a basis for dual space

Let $\mathbb V$ be finite dimensional vector space over $F$ and $\mathcal{B} = \{\alpha_1, \ldots ,\alpha_n\}$ is basis for vector space $\mathbb V$ then $\mathcal{B^*} = \{f_1, \ldots ,f_n\}$ is basis for dual space $\mathbb V^*$ such that $f_i(\alpha_{j})=\delta_{ij}$

Is the converse of this Theorem true, to explain more:

Let $\mathcal{B^*} = \{f_1, \ldots ,f_n\}$ be basis for dual space $\mathbb V^*$. Does there exist $\mathcal{B} = \{\alpha_1, \ldots ,\alpha_n\}$ such that $\mathcal{B}$ would be basis for vector space and $f_i(\alpha_{j})=\delta_{ij}$?

  • 5
    $\begingroup$ Since $\Bbb{V}$ is finite-dimensional, there's a natural isomorphism between $\Bbb{V}$ and $\Bbb{V}^{**}$. You can get the converse by applying the original theorem to $\Bbb{V}^*$ and then using that isomorphism. $\endgroup$ – Micah Jul 9 '12 at 17:54

Yes, you can prove that via the bidual space $V^{**} = \mathrm{Hom}(V^*,K)$. There is a canonical homomorphism $\iota: V \to V^{**}$ defined by $\iota(v)(f) := f(v)$ for $f \in V^{*}$. It is easy to see that $\iota$ is injective. Since $\dim V^{**} = \dim V^* = \dim V$, it is actually an isomorphism. Therefore, for every $\varphi \in V^{**}$ there exists a $v \in V$ s.t. $\varphi(f) = f(v)$ for $f \in V^*$.

Now, if $\{f_1,\ldots,f_n\}$ is a basis if $V^*$ then the elements $\{\varphi_1,\ldots,\varphi_n\}$ form a basis of $V^{**}$ where $\varphi_i(f_j) := \delta_{ij}$. By the above considerations, each $\varphi_i$ corresponds to an $\alpha_i \in V$ s.t. $f_i(\alpha_j) = \varphi_j(f_i) = \delta_{ij}$. Since $\iota: V\to V^{**}$ is an isomorphism and $\{\varphi_1,\ldots,\varphi_n\}$ is a basis of $V^{**}$ we know that $\{\alpha_1,\ldots,\alpha_n\}$ is a basis of $V$. Its dual basis is then $\{f_1,\ldots,f_n\}$ because $f_i(\alpha_j) = \delta_{ij}$.


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.