# Hilbert vs Inner Product Space

What is the difference between a Hilbert space and an Inner Product space? They both seem to be defined as simply a vector space equipped with an inner product.

Also can a metric always be defined by an inner product?

• The first is a complete inner product space. – user63181 Mar 24 '14 at 21:26

If you have a vector space $X$ with an inner product $\langle \cdot, \cdot \rangle$, this defines a norm $\|\cdot\|$ by $\|x\|=\sqrt{\langle x, x\rangle}$ (it is a good exercise to prove that this is in fact a norm). Similarly, this defines a metric, $d(x,y)=\|x-y\|$ (it is again a good exercise to prove that this is in fact a metric). This is the case for any inner product space, so yes, an inner product always defines a metric. However, not every metric is defined by an inner product!

A sequence of elements $\{x_n\}$ in $X$ is called a Cauchy sequence if $\|x_n-x_m\|\to0$ as $n,m\to\infty$. An inner product space $X$ is called a Hilbert space if it is a complete metric space, i.e. if $\{x_n\}$ is a Cauchy sequence in $X$, then there exists $x\in X$ with $\|x-x_n\|\to0$ as $n\to\infty$.

• Note that every finite dimensional vector space is complete, so for the finite dimension case inner product spaces are Hilbert spaces. – user98602 Mar 24 '14 at 21:39
• @Mike Why is that true? Unless I'm wrong (which I very easily could be) the space of rational numbers is a vector space, but it isn't complete. – user137731 Mar 24 '14 at 21:48
• @Bye_World You're very much correct! I meant to say real vector space. Thanks for correcting me. – user98602 Mar 24 '14 at 21:59

A Hilbert space is an inner product space that is complete with respect to the norm. Completeness is what differentiates the two.

Not every metric space can be defined by an inner product, for instance the space of continuous functions on $[0,1]$ with the supremum norm as its metric doesn't come from an inner product.

In particular, a Banach space (a complete normed vector space) is a Hilbert space iff its norm satisfies the parallelogram law.