We prove the exchange theorem to show well-definedness of the term "dimension" later.
In this article we will discuss the exchange lemma and Steinitz's exchange theorem. These state how a given basis of a vector space can be converted into another one by cleverly replacing some of the old basis vectors with new vector space elements. This is especially useful when you want to construct a basis that contains certain previously fixed vectors. Another consequence of the replacement theorem is the fact that linearly independent sets have a lower or equal cardinality than bases. This result is essential for the definition of the dimension of a vector space. We first prove the exchange lemma and then Steinitz's theorem.
Theorem (Exchange lemma)
Let
be a vector space over a field
and
a Basis of
. Further, let
which can be written as a linear combination
, where
. If
such that
, then
is also a Basis of
.
Proof (Exchange lemma)
Let
be the set where
has been replaced by
. We need to prove that
is a basis, as well. For this, we show that
is a generator of
and linearly independent.
Proof step:
is a generator
We know
with
and
. According to the above assumption,
and therefore it has an inverse in
.
Thus we may transform:
Because
is a basis of
, we may find for every vector
some scalars
, such that
. We now plug in the above result for
and obtain:
So the vector
has been represented as a linear combination of
and
is indeed a generator of
.
Proof step:
is linearly independent
Let
, such that
. We replace
by its representation as a linear combination of the basis elements
and obtain:
Since
is linearly independent, we have that
and
for all
.
From
and
we have
.
But this also implies
for all
.
Hence,
is linearly independent which finishes the proof.
Hint
The converse of the exchange lemma also holds true (without proof, here):
Let
be a vector space over a field
and
be a basis of
. Further, let
with the linear combination
, where
. If
such that
is also a basis of
, then we already have that
.
Next, we prove a slight modification of the exchange lemma. It shows that the lemma is "almost always" applicable. We only assume that the new basis vector
is not the zero vector:
Theorem (Exchange lemma version 2)
Let
be a vector space over a field
and
a basis of
. Further, let
. Then there is an index
such that
is also a basis of
We can thus exchange
for
.
Proof (Exchange lemma version 2)
We write
as a linear combination of vectors from
. Let also
with
.
Since
, at least one of the scalars
must be non-zero. Indeed, if we had
for all
, then
would also have to hold. So let
such that
. With this
the condition from the upper version of the exchange lemma is fulfilled.
Example
Let
be the canonical basis of
and
.
We show that
is a basis of
.
According to the exchange lemma, we can replace the vector
by
if for the (unique) linear combination
we have that:
.
We recognise that:
Thus it follows from the exchange lemma that
is a basis of
.
But we also see from this discussion that we cannot use the exchange lemma to show that
is a basis: In the linear combination of
, we have
. Therefore, we cannot apply the exchange lemma here.
We even have that
is really not a basis of
:
, so the vectors are linearly dependent, so in particular they are not a basis.
Example (Basis completion by substitution in a known basis)
We want to complete the two vectors
to a basis
of
.
First we should check whether the two vectors are linearly independent, but this is easy to see because for
there must be
, which then means that
as well.
We now want to exchange
for two basis vectors of a known basis of
using the exchange lemma. For the
we simply take the canonical basis
.
By repeatedly applying the exchange lemma, the vectors are exchanged one after the other. Here we proceed as follows:
The vector
is represented as a linear combination of the vectors
.
According to the exchange lemma, we can exchange any basis vector
, since for every
the associated scalars are
.
So, for example, if we exchange
for
, then we get that
is a basis of
.
We now repeat the above procedure for
with the basis
.
Again, we may exchange any of the vectors
, but exchanging
is not purposeful, so we exchange
. Thus the vectors
form a basis of
. So in total we have completed the linearly independent vectors
with
to a basis of
.
Theorem (Steinitz's theorem)
Let
be an
-element basis of the
-vector space
and let
be a
-element set of linearly independent vectors.
Then, we have that
and there is a certain choice of
vectors from the basis
that can be replaced by the vectors
such that the replacement results in a new basis
of the
-vector space
.
More precisely, after renumbering the indices (if necessary) we can write
Proof (Steinitz's theorem)
We prove the theorem by induction over k.
Proof step: Induction base 
Let
be a basis of
and let
be linearly independent, i.e.
, then it follows with above exchange lemma, that
for a given
is a basis of
.
Now we rename
to
.
It then follows that
is a basis. This is the base case of our induction.
Proof step: Induction step 
Let
be linearly independent. By the induction assumption, we have that
, from which we want to infer
. Suppose there was
. Then since
and
we have that
. As
is linearly independent and
, we can replace
with
by induction assumption. So we get that
is a basis of
.
We can also represent
as a linear combination in
. Let
, with
. Then we have:
But this is a contradiction to the linear independence of
. So
must also hold true.
It remains to show that after any renaming of
is a basis of
. By induction assumption,
is a basis of
. In this case, the indices of
may have been interchanged. We write
as a linear combination of
.
Let for this
, with
. Note that these are not necessarily the same
as before. Suppose there was
for all
. Then
would hold, which would be a contradiction to the linear independence of
. So, let
with
. By the exchange lemma,
is a basis of
.
Finally, we rename
to
.
Then
is a basis of
, which concludes the proof.