# While proving that every vector space has a basis, why are only finite linear combinations used in the proof?

Statement: Every vector space has a basis

Standard Proof:It is observed that a maximal linearly independent set is a basis. Let $\mathscr{Y}$ be a chain of linearly independent subsets of a vector space $\mathscr{V}$. The union of such a set can serve as an upper bound for it.To apply Zorn's lemma,we have to check whether the union is linearly independent? Well, if $t_1,\dots,t_n$ belong to the union, then each $t_i$ belongs to some linearly independent set $L_i\in \mathscr{Y}$. Because $\mathscr{Y}$ is a chain, one of these sets $L_i$ contains all the others. If that is $L_j$, then the linear independence of $L_j$ implies that no non-trivial linear combination of $t_1,\dots,t_n$ can be zero, which proves that the union of the sets in $\mathscr{Y}$ is linearly independent. Therefore, by Zorn’s lemma, there is a maximal linearly independent set and hence a basis.

My question: Why are we using only finite linear combinations to show that the union is linearly independent. Surely, if the union is infinite,then there do exist many infinite linear combinations of elements of the union, which cannot be proven to be linearly independent by the same reasoning. I suspect, that perhaps we are not concerned with infinite linear combinations due to issues of convergence, but I'm not sure.

-
Linear dependence, basis, are precisely about finite linear combinations. In a restricted number of cases, some infinite sums make sense, but that is another subject. –  André Nicolas Jul 2 '12 at 12:59
There needs to be a way of ascribing a meaning to infinite sums. –  copper.hat Jul 2 '12 at 13:00

The definition of linear independence is that any finite linear relation is trivial.

Vector spaces in general do not have any concept of an infinite sum at all. For those vector space where the usual concept of an infinite sum of reals can be generalized, one may speak of a different kind of span/basis where one allows infinite linear combinations in addition to finite ones. That gives rise to a separate concept, different from the usual kind of linear-combinations basis.

When one needs to distinguish between the different notions of basis, an ordinary basis that works by finite linear combinations is called a "Hamel basis" or "algebraic basis", and one that needs infinite linear combinations to span everything is called a "Schauder basis" (though strictly speaking the latter name implies some additional conditions).

-
Thanks. That really helped. –  user33109 Jul 2 '12 at 13:46

A vector space has no notion of convergence or order. Since these are required for assigning meaning to an infinite sum, we are restricted to finite sums.

In the study of normed linear spaces, the notion of Schauder basis is an object of study.

-