1

I'm trying to rebase my know how in linear algebra, restart from scratch to get a more formal and useful set of definitions to help me with computer programming stuff .

One of the first concepts is a vector space, which is a space with :

  • a field $F$ of scalars
  • a group of vectors $V$
  • a vector sum and a vector multiplication
    • said operations are commutative and associative
    • the sum has $0$ as identity
    • the multiplication has $1$ as identity

A vector space is also an algebraic structure .

You then use vector spaces to define Hilbert spaces .

An Hilbert space is a vector space where you can perform an inner product, in particular a scalar product is needed ( aka dot product ) .

I'm puzzled by the introduction of this scalar product and why is needed, the reported consequences of this are that :

  • you can talk about:
    • angles
    • distances
    • orthogonality of lines
  • limits exist, and this is good so you can apply calculus to Hilbert spaces
  • and because limits exist ( I guess that this means that everything always converge to a point ) you can pretty much always express an Hilbert space ( or elements of an Hilbert space (?) ) with a series, a series of elements from your field ( numbers ) or a series of functions

It's not explicitly stated in any resource I found but I think that Hilbert spaces are not algebraic structure as vector spaces are, you lose this property .

Assuming that I haven't made any mistakes, is this a list of consequences from just having a scalar product defined for a vector space ?

What is a vector space without a scalar product defined ? What's a vector space that is not an Hilbert space ? How it looks like ?

I'm really having some troubles in visualizing this with some geometries or just some graphs .

Caleb Stanford
  • 45,048
  • 8
  • 70
  • 160
user2485710
  • 505
  • 4
  • 14
  • Posts related to part about inner product spaces: [Is there a vector space that cannot be an inner product space?](http://math.stackexchange.com/questions/247425/is-there-a-vector-space-that-cannot-be-an-inner-product-space) and [Does every vector space have an inner product?](http://math.stackexchange.com/questions/976740/does-every-vector-space-have-an-inner-product). – Martin Sleziak Jan 18 '15 at 16:13
  • About Hilbert spaces: [Can all real/complex vector spaces be equipped with a Hilbert space structure?](http://math.stackexchange.com/questions/427834/can-all-real-complex-vector-spaces-be-equipped-with-a-hilbert-space-structure) (Which is related to [Let $X$ be an infinite dimensional Banach space. Prove that every Hamel basis of $X$ is uncountable.](http://math.stackexchange.com/questions/217516/let-x-be-an-infinite-dimensional-banach-space-prove-that-every-hamel-basis-of) and [Cardinality of a Hamel basis](http://math.stackexchange.com/questions/141535/cardinality-of-a-hamel-basis)). – Martin Sleziak Jan 18 '15 at 16:17

1 Answers1

1

Each Hilbert space $(X, <\cdot,\cdot>)$ is also a normed space $(X, ||\cdot||)$ with norm $||x|| := \sqrt {<x,x>}, x\in X$.

Concerning the opposite direction: If $(X, ||\cdot||)$ is a normed vector space, then its norm must satisfy the parallelogram law $||x+y||^2 + ||x-y||^2=2||x||^2 + 2||y||^2$ in order to derive from an inner product.

Not every normed space is also a Hilbert space. The p-norm on $X=\mathbb R^n, n>1$, defined as $||x||_p:=(\sum_{i=1,...,n} |x_i|^p)^\frac{1}{p}, p\neq 2,$ does not derive from an inner product.

Jo Wehler
  • 2,205
  • 16
  • 30