I'm beginning a self-study of functional analysis, and I seem to have come to a halt trying to solve the first problem in the first problem set, and was wondering if someone could give me a pointer.

The problem asks to show that, while the property $1\mathbf{a}=\mathbf{a}$ must be included in the axioms of a vector space, it is a property that can be derived from the axioms of a normed vector space.

The axioms for a vector space $V$ over the field $\mathbf{R}$ given in the book are as follow:

  1. $V$ is a group.
  2. For each $\alpha\in\mathbf{R}$ and $\mathbf{a}\in V$, $\alpha\mathbf{a}\in V$
  3. $\alpha(\mathbf{a}+\mathbf{b})=\alpha\mathbf{a}+\alpha\mathbf{b}$
  4. $(\alpha+\beta)\mathbf{a}=\alpha\mathbf{a}+\beta\mathbf{a}$
  5. $\alpha(\beta\mathbf{a})=(\alpha\beta)\mathbf{a}$
  6. $1\mathbf{a}=\mathbf{a}$

The axioms for a normed vector space are 1--5 plus the additional axiom that there is a map $\Vert\cdot\Vert:V\to\mathbf{R}$ that satisfies all the properties of a norm.

The problem formally stated now is to show that 1--5 do not imply 6, but 1--5 together with the norm do imply 6.

This problem seems similar to the question posed here, and from the discussion of that question it seems an easy example of a vector space satisfying 1--5 but not 6 would be one where scalar multiplication always yields the zero vector. This cannot be the case in a normed vector space, because for $\mathbf{a}\neq0$ and $\alpha\neq0$ we know $\Vert\mathbf{a}\Vert\neq0$, and a norm satisfies $\Vert\alpha\mathbf{a}\Vert=\vert\alpha\vert\Vert\mathbf{a}\Vert\neq0$, but we also have $\Vert\alpha\mathbf{a}\Vert=\Vert\mathbf{0}\Vert=0$ from our definition of scalar multiplication, which is a contradiction.

It is trivial that $1\mathbf{a}=\mathbf{a}$ if every vector in $V$ can be written as a scalar multiple of a vector in $V$, since then $\mathbf{a}=\alpha\mathbf{b}=(1\alpha)\mathbf{b}=1(\alpha\mathbf{b})=1\mathbf{a}$, but while the norm has ruled out the strange case of scalar multiplication that means only $\mathbf{0}$ can be written as a scalar multiple of a vector, I haven't been able to convince myself that other cases where there exist vectors that are not scalar multiples of a vector have been ruled out.

Can anyone offer a pointer for how to show every vector in a normed vector space can be written as a scalar multiple of a vector in the space, or perhaps point me to a more natural way of approaching this problem?



Nice question, this took me more thought than I expected.

Consider $a\in V$. Note that $0=\|1a-1a\|=\|1a-(1)(1a)\|=\|(1)(a-1a)\|=\|a-1a\|$.

Thus $a=1a$.


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.