There is a theorem which says that if $V$ is a vector space over a field $\mathbb{K}$ and $V$ admits an infinite linearly independent set, then any two bases $B$ and $B'$ have the same cardinality. So I wanted to prove this theorem. The question is followed by a hint which says that the following lemma can be applied to prove this theorem.

Lemma: Let $A\ne \emptyset \ne B$ be two sets. A map $\phi: B\to C$ is countable-to-one if for every $c\in C$, $\phi^{-1}(\{c\})$ is countable. If such a map exists then $|B| \le \aleph_0|C|$.

The idea was to apply this lemma to the collections of finite sets of $B$ and $B'$, i.e. $\mathcal{F}(B)$ and $\mathcal{F}(B')$, which is quite straightforward, except that $\mathcal{F}(B)$ may well be uncountable.

So I asked my professor for a hint and here's what he said:

Given $E\in\mathcal{F}(B)$, there is a minimal finite set $f(E)$ of elements of $B'$ for which $E$ is contained in span$(f(E))$.

He also added that this is a simple fact from linear algebra. But I didn't recognize this fact. Can someone please explain the concept from linear algebra above in simpler terms?

Namely, am I correct in re-interpreting this statement as follows?:

Given $E\in\mathcal{F}(B)$, there is a bijective function $f$ on $E$ which takes elements of $B$ into elements of $B'$ such that $E$ is spanned by $f(E)$ (that is, by a minimal finite subset of $B'$).

Anyway, this statement doesn't seem to be a trivial concept from linear algebra, and I don't think I've ever heard about it.


The trivial fact that your professor alluded too is basically the definition of the span of an infinite set of vectors. If $B$ is an infinite set of vectors then $\text{span} \ B$ is by definition the set of vectors that can be written as a linear combination of some finite subset of vectors in $B$.

The span of a basis is the entire space so given any one vector $v \in B$ there is a finite subset of $B'$ such that $v$ is in the span of that subset. Thus given any finite subset $E \in \mathcal F(B)$ there is a finite subset of $B'$ whose span contains $E$.

As for your re-interpretation: If you think about the one element subsets of $B$ your re-interpretation seems to say that every element of $B$ is a scalar multiple of an element in $B'$, and this is not necessarily true so your re-interpretation looks incorrect to me.

  • $\begingroup$ Do you mean it is necessarily true that if $B$ is an infinite set of vectors then it can be spanned by a finite basis? Then, I think, infinite bases would not exist. $\endgroup$ – sequence Sep 23 '17 at 0:44
  • $\begingroup$ No, this is only true for finite sets of vectors. $\endgroup$ – Jim Sep 23 '17 at 1:41
  • 1
    $\begingroup$ I did not say that. I said that for each vector $v$ in the span of $B$ there is a finite set of vectors from $B$ such that $v$ is a linear combination of the vectors in this finite set, but I did not say that you can use the same finite set of vectors each time. $\endgroup$ – Jim Sep 24 '17 at 3:27
  • 1
    $\begingroup$ $\mathbb R$ does not span $\mathbb Q$, it's not even contained in $\mathbb Q$. Generally if $V$ is a vector space over the coefficient field $k$ and $S$ is a set of vectors in $V$ then $\text{span}_k(S)$ is the subspace of $V$ whose elements are vectors that can be written in the form $\sum_{i = 1}^na_iv_i$ for some choice of a positive integer $n$, $n$ coefficients $a_i \in k$, and $n$ vectors $v_i \in S$. It $\endgroup$ – Jim Sep 27 '17 at 22:56
  • 1
    $\begingroup$ And as for infinitely many rationals that add to a real, yes you can write a real number as a limit of an infinite sequence of sums of rational numbers, but the definition of a span in linear algebra is that you are only allowed to have finitely many vectors. $\endgroup$ – Jim Sep 27 '17 at 23:07

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.