MATH1012 Mathematical Theory and

Methods

Week 4

1 / 56
Determinants

A number associated with any square matrix:

 
a b
det = ad − bc
c d

Alternative, but common, notation

a b
= ad − bc
c d

2 / 56
Inverses

Somehow related to inverses

 −1  
a b 1 d −b
=
c d ad − bc −c a

3 / 56
Three-by-three

What is  
2 1 −1
det  0 1 1 ?
1 1 −1
Defined recursively using cofactors.
     
2 1 −1 2 1 −1 2 1 −1
 0 1 1   0 1 1   0 1 1 
1 1 −1 1 1 −1 1 1 −1

4 / 56
Technically

Let A be an n × n matrix.
▶ if n = 1, then det A = a11 .
▶ if n > 1, then
n
X
det A = (−1)1+j a1j det(A[1, j]),
j=1

where A[i, j] is the matrix obtained from A by deleting the

ith row and jth column.

Expanding along the first row.

5 / 56
Any row or column

 
2 1 −1
det  0 1 1 
1 1 −1

     
2 1 −1 2 1 −1 2 1 −1
 0 1 1   0 1 1   0 1 1 
1 1 −1 1 1 −1 1 1 −1

6 / 56
The pattern

Expanding requires using this pattern:

 
+ − +
 − + − 
+ − +

+ − + − + −
 
 − + − + − + 
+ − + − + −
 
 
− + − + − +
 
 
 + − + − + − 
− + − + − +

7 / 56
Pick the right row/column
 
1 0 0 −1 2

 0 1 0 2 1 

det 
 1 0 0 −1 1 

 1 0 0 0 2 
1 2 1 5 7

8 / 56
First properties

▶ det(AT ) =

▶ if A has a row of zeroes then det(A) =

9 / 56
Upper triangular

If A is upper (or lower) triangular, then det(A) =

 
1 −2 6 −1 2

 0 3 5 2 1 

det 
 0 0 2 −1 1 

 0 0 0 5 2 
0 0 0 0 7

10 / 56
Expanding method doesn’t scale

Computing the determinant of one n × n matrix requires

computing the determinant of n matrices of order n − 1.

Computing the determinant of 100 × 100 matrix in this fashion

would take longer than the lifetime of the universe.

Coming to the rescue is our trusty friend — row-reduction.

11 / 56
First ingredient

A square matrix in row echelon form is upper triangular

   
2 0 1 3 1 2 0 3
 0 0 2 1   0 2 −1 1 
   
 0 0 0 2   0 0 −1 2 
0 0 0 0 0 0 0 2

Hence finding the determinant is easy for matrices in

row-echelon form.

12 / 56
Second ingredient

EROs do change the determinant, but only in predictable ways

If A′ is obtained from A by doing an ERO then

▶ det A′ = − det A if the ERO is Type 1: Ri ←→ Rj

 
0 −1 6
det  1 3 5 =
6 8 2

13 / 56
EROs
If A′ is obtained from A by doing an ERO then

▶ det A′ = α det A if the ERO is Type 2: Ri ←− αRi

 
1 3 5
det  0 −1 6  =
6 8 2

▶ det A′ = det A if the ERO is Type 3: Ri ←− Ri + αRj

 
1 3 5
det  0 −1 6  =
3 4 1

14 / 56
Technique

▶ Row reduce via sequence of EROs

▶ Keep track of total cumulative change in determinant

▶ Find determinant of row reduced matrix

▶ Figure out what original determinant is

15 / 56
Common Mistake (Warning!)

▶ Row-reduce A to A′
▶ Find det A′
▶ Write this down as the answer

16 / 56
Example

 
1 2 1
det  2 4 3  = d
3 1 0

17 / 56
Example

 
1 0 1 y
 0 0 0 y 
det 
 0 −1

0 1 
1 −y 2 0 y

18 / 56
Properties of determinants

▶ det(αA) =

▶ A is invertible if and only if det A ̸= 0.

19 / 56
Determinant is multiplicative

det(AB) = det(A) × det(B)

As a consequence,

▶ det(AB) = det(BA)
▶ det(Ak ) =
▶ det A−1 =

20 / 56
Invertible matrix theorem

An n × n matrix A is invertible if and only if

▶ It has rank n
▶ It has nullity 0
▶ It has non-zero determinant
▶ Its rows are linearly independent
▶ Its columns are linearly independent
▶ Its row echelon form has n pivots (leading entries)
▶ Its reduced row echelon form is In
▶ The equation Ax = 0 has a unique solution
▶ Its rows are a basis for Rn
▶ Its columns are a basis for Rn

21 / 56
So far

▶ Systems of linear equations

▶ Vector spaces R2 , R3 , . . ., Rn
▶ Subspaces have bases and a dimension
▶ Matrix addition and multiplication
▶ row/column/null space
▶ Rank-nullity theorem
▶ Matrix inverses
▶ Determinants

Now we start Chapter 4: Linear transformations.

22 / 56
Function terminology
A function has a domain A and a codomain B

and maps each a ∈ A to exactly one element in B.

23 / 56
Functions between vector spaces

Define f : R2 → R2 by

f ((x, y)) = (x + 1, x + y)

24 / 56
Linear Transformations

A linear transformation is a function f : Rn → Rm such that

▶ f (u + v) = f (u) + f (v) for all vectors u, v ∈ Rn

▶ f (αu) = αf (u) for all vector u ∈ Rn and scalar α ∈ R

25 / 56
Common Task

Given a function f , decide whether or not it is a linear

transformation.

▶ If you think that the answer is “No”

Present a counterexample — that is, two vectors v and w
where f (v + w) ̸= f (v) + f (w).
▶ If you think that the answer is “Yes”
Give a symbolic proof that works for arbitrary vectors.

26 / 56
Example
Are the following functions linear transformations?
▶ f : R2 → R2 : (x, y) 7→ (x + y, x + y)

▶ g : R2 → R2 : (x, y) 7→ (x2 , y 2 )

▶ h : R3 → R3 : (x, y, z) 7→ (x, y + z, 0)

27 / 56
Property

If f : Rn → Rm is a linear transformation then

f (0) = 0.

Proof:

28 / 56
Kernel

If f : Rn → Rm , then the kernel of f is the vectors in the

domain Rn that are mapped to 0.

29 / 56
Kernel

What is the kernel of f (x, y) = (2x − y, 2y − 4x)?

domain codomain

Exercise: what is the kernel of g(x, y) = (x + y, x − y)?

30 / 56
Subspace

If f : Rn → Rm is a linear transformation, then the kernel of f

is a subspace of Rn .
Proof:

31 / 56
Range
If f : Rn → Rm , then the range of f is the set of vectors

{f (u) : u ∈ Rn }.

This is a set of vectors in the codomain of f , namely Rm .

32 / 56
Range

What is the range of f (x, y) = (2x − y, 2y − 4x)?

domain codomain

Exercise: what is the range of g(x, y) = (x + y, x − y)?

33 / 56
Subspace

If f : Rn → Rm is a linear transformation, then the range of f is

a subspace of Rm

34 / 56
Determining linear transformation

A linear transformation is determined by its action on a basis of

the domain.

Suppose f : R2 → R3 is a linear transformation such that

f (1, 0) = (1, 1, 1) and f (0, 1) = (0, 0, 1)

▶ f (1, 1) =

▶ f (2, 4) =

▶ f (x, y) =

35 / 56
Consequence

The range of a linear transformation f : Rn → Rm is the span of

the images of any basis of Rn .

What is the range of f (x, y) = (2x − y, 2y − 4x)?

36 / 56
Finding bases

The range and kernel of a linear transformation are subspaces.

So we will want to find a basis for each of these subspaces.

Fortunately, we already have the tools and techniques needed!

37 / 56
Basis for kernel

To find a basis for the kernel, we must solve f (v) = 0

What is the kernel of

f : R3 → R3
given by
f ((x, y, z)) = (x − y, y − z, z − x)?

38 / 56
Basis for range

If f : R3 → Rm is linear then every vector in its range is a linear

combination of

{f ((1, 0, 0)), f ((0, 1, 0)), f ((0, 0, 1))}

So these vectors form a spanning set for the range, but we need
a basis.

39 / 56
Example

Let f : R3 → R2 be given by f (x, y, z) = (x + y, y + z)

Find a basis for the kernel and range of f .

40 / 56
Matrix multiplication

Let A be an arbitrary m × n matrix and define f : Rn → Rm by

f (v) = Av

where v is an n × 1 column vector.

Then f is a linear transformation.

Our aim is to show that every linear transformation is just

matrix multiplication (in disguise).

41 / 56
Matrix of transformation

To represent a linear transformation as a matrix, need to

specify a basis for the domain and codomain.

▶ Linear transformation f : Rn → Rm
▶ B = {v1 , v2 , . . . , vn } basis for Rn
▶ C = {w1 , w2 , . . . , wm } basis for Rm

The matrix [f ]CB is an m × n whose i’th column is the

C-coordinate vector of f (vi ).

42 / 56
Example

Find the matrix of the reflection in R2 through the line

x + y = 0 with respect to the standard basis S (for both the
domain and codomain)

43 / 56
Example

Let f : R2 → R2 be given by

f (x, y) = (x + y, x − y)

and let B = {(1, 1), (2, 1)} and C = {(1, 1), (1, 0)}.

44 / 56
Use of matrix

If fCB is the matrix of linear transformation f , then

fCB (v)B = (f (v))C .

In other words, multiplying by fCB takes a B-coordinate

vector, applies the linear transformation f , and expresses the
result in C-coordinates.

45 / 56
Example
Find the matrix of g w.r.t bases C, D where

C = {(1, 1), (1, 0)}

D = {(1, 0), (0, 1)}

and g : R2 → R2 is defined by

g(x, y) = (x + 2y, 2x + y)

46 / 56
Exercise

Let f : R3 → R2 be given by f (x, y, z) = (x + y, y + z)

B = {(1, 1, 1), (1, 1, 0), (1, 0, 0)} C = {(1, 1), (1, −1)}

47 / 56
Example

Find the matrix of the identity transformation in R3 w.r.t any

basis B (for domain and codomain)

48 / 56
Back to kernel and range

Suppose that A = fSS for some linear transformation f .

▶ What is the range of f ?

▶ What is the kernel of f ?

49 / 56
Another rank-nullity theorem

If f : Rn → Rm is a linear transformation, then

dim(range(f )) + dim(ker(f )) = n

50 / 56
Composition of functions

If f : Rℓ → Rm and g : Rm → Rn then

g ◦ f : Rℓ → Rn

is defined by
(g ◦ f )(v) = g(f (v))

Then g ◦ f is linear.

51 / 56
Matrix of composition

Consider f (x, y) = (x + y, x − y) and g(x, y) = (x + 2y, y + 2x)

with bases

B = {(1, 1), (2, 1)}

C = {(1, 1), (1, 0)}
D = {(1, 0), (0, 1)}

and found
   
0 1 3 1
[f ]CB = [g]DC =
2 2 3 2

What is the matrix of g ◦ f with respect to B and D?

52 / 56
Working

f g
(1, 1) −→ −→

f g
(2, 1) −→ −→

= (1, 0) + (0, 1)
= (1, 0) + (0, 1)

So the matrix is
 
[g ◦ f ]DB =

53 / 56
Matrix of composition

The matrix of a composition of two linear transformations is the

product of the matrices.

54 / 56
Inverse function

An arbitrary function f : A → B is called invertible if there is a

function g : B −→ A such that f −1 (g(x)) = x for all x ∈ A.

If this function exists it is denoted f −1 .

For example, suppose f : R2 → R2 is given by

f (x, y) = (x + y, x − y)

55 / 56
Inverse of a linear transformation

If the linear transformation f : Rn → Rn is invertible, then

▶ The map f −1 is a linear transformation
▶ The matrix [f −1 ]SS is

56 / 56