Throughout this Chapter we will use the letter to denote any field; but usually, in exercises and applications, it will mean either or . The notion of a linear transformation was introduced in MATH105 as a function from to . We will restate the definition here, in terms of arbitrary vector spaces.
Let and be vector spaces over the same field . A function is called a linear transformation if it satisfies the following two conditions:
for any ,
for any and .
Here is the domain of , and is the codomain of .
Let for a field . Then the function defined as follows is a linear transformation:
for all . Here we consider elements of as column vectors.
As we have seen in MATH105 for , every linear transformation can be expressed as for some matrix .
[Caution: The phrase “Linear transformation” is used differently in MATH230. In that module the functions of the form , where is a non-zero vector are also considered “linear transformations” (unlike this module). Also, other sources sometimes prefer the name “linear map” or “vector space morphism”.]
Let be defined by . Find a matrix such that .
Solution: Write for the standard basis of , and to avoid duplicating notation, we write for the standard basis of . Then we compute that
Finally, create by taking the columns to be the coordinates of with respect to the standard basis of . So .
The above example should be familiar from MATH105. It makes use of the standard basis of . The following generalization allows for non-standard bases as well.
Let , be vector spaces over the same field , and assume:
is a basis of , and
is basis of . If is a linear transformation, then the matrix of with domain basis and codomain basis is constructed as follows:
In other words, the columns are the coordinates of with respect to the basis . In the case when we also simply write:
If no basis is specified, then the matrix of a linear transformation is defined as above, but using the standard basis for and , as in Example 4.3.
Let be defined by , and let be a basis for . Compute the matrix of with respect to the basis in the domain and codomain.
Solution: We compute the coordinates as as follows:
Using these coordinates as the column vectors, we find .
Consider the linear transformation defined by . Prove that the matrix of with respect to the basis in both the domain and codomain is:
[End of Exercise]
Let be a linear transformation, and bases for and respectively. Then for any vector we have
Recall that is the column vector of coordinates of with respect to , and is the column vector of coordinates of with respect to .
In other words, the matrix transforms the coordinate vector to . The following exercise verifies this theorem is some specific cases.
Let , and , and let be the standard basis of . For each of the following bases, compute and . Hence verify Theorem 4.7 for the vector in each case:
is the standard basis of .
.
.
[End of Exercise]
If , , and are all bases of , and are linear transformations, then we have
The proof repeatedly uses Theorem 4.7. For any we have:
But if for all vectors , then . The result follows. ∎
It’s as if the neighbouring “”s cancel each other out. This is the reason for writing the notation as it is, and is a good trick for manipulating these matrices.