Prove that every vector space has a basis. I am going to use Zorn's lemma for this also here is a necessary definition regarding totally ordered subsets: one element will be contained in the other.

proof: A basis for $V$ is a maximally linearly independent set. Let $X$ be equal to all linearly independent subsets of $V$. On, $X$ define $A\leq B$ means $A\subset B$. Let $S$ be a totally ordered subset of $X$, let $T = \bigcup_{A_{s}\in S} A_s$ then $A_s\subset T$ for all $s\in S$. Does $T\in S$? i.e. is $T$ a linearly independent set?. Let $$O = \alpha_1 v_1 + \alpha_2 v_2 + \ldots + \alpha_n v_n \ \ \text{where} \ \ v_i\in T = \bigcup_{A_{s}\in S} A_s$$ Then there exists $S_i$ such that for all $i\in A_{s_{i}}$ all $i$ $A_{s_{1}},A_{s_{2}},\ldots, A_{s_{n}}$ (these will have one contained in the other then we would take the larger). So, there exists $A_{s_{k}}$ such that $A_{s_{k}}\supset A_{s_{i}}$ all $i$ just compare two at a time. Therefore, $V_i \in A_{s_{k}}$ all $i$, but $A_{s_{k}}$ is a linearly independent set. So, $\alpha_i = 0$ all $i$. Hence, $T\in X$ so $T$ is an upper bound for $S$. Hence every totally ordered subset of $X$ has an upper bound, hence by Zorn's lemma, there exists a maximal element and this would be the basis.

I am not sure if my proof is right, I apologize if it is a bit messy. If it is right, any suggestions on cleaning it up would be greatly appreciated.


The proof is very hard to understand. Here is why:

  1. Either the elements of $S$ are $s$ or $A_s$, it's confusing to mix those apart.

  2. You have a confusing typo, when you write "Does $T\in S$?" you mean $T\in X$.

  3. $0$ and $O$ are two different things.

  4. $i$ seems to be some index, but then you say that $i\in A_{s_i}$ which makes absolutely no sense. Is it some finite indexing? Is it a vector in $A_s$? What is it?

  5. Where did $V_i$ come from? Perhaps you meant $v_i$? Remember that mathematics is case sensitive.

Other than that, if I understood the underlying idea you were trying to use, this is fine. Here's a simpler and cleaner version of what you wrote:

Let $X$ be all the linearly independent subsets of $V$, and let $\leq$ be the $\subseteq$ order on $X$. Let $S$ be a totally ordered subset of $X$, then we claim that $T=\bigcup S$($=\bigcup_{A\in S}A$) is an upper bound of $S$ and $T\in X$.

$T$ is a linearly independent set, since if $\alpha_1\cdot v_1+\ldots+\alpha_n\cdot v_n=0$ for some scalars $\alpha_i$ and $v_i\in T$, there are some $A_i\in S$ such that $v_i\in S_i$. But $S$ is linearly ordered so there is some $j$ such that $v_i\in A_j$ for all $1\leq i\leq n$. And because $A_j$ is linearly independent it has to be that $\alpha_i=0$ for all $i$. Therefore $T$ is linearly independent and $T\in X$.

It is an upper bound of $S$ since for every $A\in S$, it is true that $A\subseteq T$, so $A\leq T$.

We have shown that given an arbitrary totally ordered subset of $X$, it has an upper bound in $X$, so by Zorn's lemma there is a maximal linearly independent set, which is a basis. $\square$


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.