Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

I am working through Theorem 6 in Chapter 8 of Hoffman and Kunze. (The section is Linear Functionals and Adjoints.)

The statement is

Let $V$ be a finite-dimensional inner product space, and $f$ a linear functional on $V$. Then there exists a unique vector $\beta$ in $V$ such that $f(\alpha) = (\alpha | \beta)$ for all $\alpha$ in $V$.

(EDIT: The following is written in terms of a real vector space.)

I am fine with the fact that this essentially follows from what $f$ does to the elements of a basis for $V$, in other words:

For an arbitrary* $\mathcal{B} = \{e_1, \ldots, e_{n} \}$ (where $n = \dim V\;\;$) we can assign $\xi_{i} = f(e_{i})$ for $i=1,\ldots, n$. Then any $\alpha \in V$ can be written as $\alpha = x_{1}e_{1} + \cdots + x_{n}e_{n}$ so that $$\begin{align} f(\alpha) &= f(x_{1}e_{1} + \cdots + x_{n}e_{n}) \\ \\ &= x_{1}f(e_{1}) + \cdots + x_{n}f(e_{n}) \\ \\ &= x_{1} \xi_{1} + \cdots + x_{n}\xi_{n}.\end{align}$$

At this point we can identify the $\xi_{i}$'s as being independent of $\alpha \in V$, and only having been dependent on our choice of a basis. With this, we can simply assert that $\beta = (\xi_{1}, \ldots, \xi_{n})$ (and is in fact unique).

Now, this is apparently not the whole story! *According to H. and K., the coordinates of our fixed vector $\beta$ are determined by the action of $f$ on an orthonormal basis for $V$ rather than an arbitrary basis.

Question: What role does the orthogonality (or even the 'orthonormality') of the basis have in this argument? It seems to work just fine for an arbitrary basis as far as I can tell.

share|cite|improve this question
up vote 3 down vote accepted

If the basis $\{e_1,\ldots,e_n\}$ is not orthonormal, then the inner product of the vectors $\alpha = x_1e_1 + \cdots + x_n e_n$ and $\beta = \xi_1 e_1 + \cdots + \xi_n e_n$ is not equal to $x_1\xi_1+\cdots+x_n\xi_n$. Thus, while it is true that $f(\alpha) = x_1\xi_1+\cdots+x_n\xi_n$, it is not true that $f(\alpha)=(\alpha|\beta)$.

share|cite|improve this answer
Jim, thank you for pointing this out. Now I recognize that without orthogonality of the basis, $\alpha \neq (x_{1},\ldots,x_{n})$, and without unit length of each basis element, the products $\| e_{i}\|x_{i}\xi_{i} \neq x_{i}\xi_{i}$. – Tom Stephens Sep 5 '10 at 4:21
I have to unaccept this answer because I am no longer convinced of the claim that $f(\alpha)\neq(\alpha | \beta)$ in the absence of an orthonormal basis. Could you expand on that a bit? – Tom Stephens Sep 7 '10 at 20:18
In general, $(\alpha|\beta) = \sum_{i=1}^n \sum_{j=1}^n x_i\xi_j(e_i|e_j)$ by the linearity of the inner product. If the basis is orthonormal, then $(e_i|e_j)$ is 1 for $i=j$ and 0 otherwise, in which case the sum reduces to the usual formula for the inner product. – Jim Belk Sep 8 '10 at 5:10

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.