MA102 Linear Algebra

Lectures-11 : Linear Transformations

RA, RKS, MGPP, BD

Department of Mathematics
IIT Guwahati

March-June 2023

1 / 15
Linear Transformations

Topics:
Linear Transformation
Kernel and Range
Rank-Nullity Theorem
Invertible Linear Transformations

2 / 15
Linear Transformations
Definition: Let V and W be vector spaces over the field F. A map T : V −→ W is
called a linear transformation (LT)

3 / 15
Linear Transformations
Definition: Let V and W be vector spaces over the field F. A map T : V −→ W is
called a linear transformation (LT) if for all u, v ∈ V and α ∈ F

T (u + v) = T (u) + T (v) and T (αv) = α T (v).

3 / 15
Linear Transformations
Definition: Let V and W be vector spaces over the field F. A map T : V −→ W is
called a linear transformation (LT) if for all u, v ∈ V and α ∈ F

T (u + v) = T (u) + T (v) and T (αv) = α T (v).

Equivalently, for all u, v ∈ V and α, β ∈ F

T (αu + βv) = αT (u) + βT (v).

3 / 15
Linear Transformations
Definition: Let V and W be vector spaces over the field F. A map T : V −→ W is
called a linear transformation (LT) if for all u, v ∈ V and α ∈ F

T (u + v) = T (u) + T (v) and T (αv) = α T (v).

Equivalently, for all u, v ∈ V and α, β ∈ F

T (αu + βv) = αT (u) + βT (v).

Notation: Often we write T (u) simply as T u.

3 / 15
Linear Transformations
Definition: Let V and W be vector spaces over the field F. A map T : V −→ W is
called a linear transformation (LT) if for all u, v ∈ V and α ∈ F

T (u + v) = T (u) + T (v) and T (αv) = α T (v).

Equivalently, for all u, v ∈ V and α, β ∈ F

T (αu + βv) = αT (u) + βT (v).

Notation: Often we write T (u) simply as T u.

The zero transformation: 0 : V −→ W given by 0(v) = 0 for all v ∈ V.

3 / 15
Linear Transformations
Definition: Let V and W be vector spaces over the field F. A map T : V −→ W is
called a linear transformation (LT) if for all u, v ∈ V and α ∈ F

T (u + v) = T (u) + T (v) and T (αv) = α T (v).

Equivalently, for all u, v ∈ V and α, β ∈ F

T (αu + βv) = αT (u) + βT (v).

Notation: Often we write T (u) simply as T u.

The zero transformation: 0 : V −→ W given by 0(v) = 0 for all v ∈ V.

The identity transformation: I : V −→ V given by I (v) = v for all v ∈ V.

3 / 15
Linear transformations from Rn to Rm
• Let A ∈ Rm×n . Define TA : Rn −→ Rm by TA (x) := Ax for all x ∈ Rn .

4 / 15
Linear transformations from Rn to Rm
• Let A ∈ Rm×n . Define TA : Rn −→ Rm by TA (x) := Ax for all x ∈ Rn . Then TA is
linear.

4 / 15
Linear transformations from Rn to Rm
• Let A ∈ Rm×n . Define TA : Rn −→ Rm by TA (x) := Ax for all x ∈ Rn . Then TA is
linear. Indeed, TA (αx + βy) = A(αx + βy) = αAx + βAy = αTA (x) + βTA (y).

4 / 15
Linear transformations from Rn to Rm
• Let A ∈ Rm×n . Define TA : Rn −→ Rm by TA (x) := Ax for all x ∈ Rn . Then TA is
linear. Indeed, TA (αx + βy) = A(αx + βy) = αAx + βAy = αTA (x) + βTA (y).

• If T : Rn −→ Rm is a linear transformation, then there exists A ∈ Rm×n such that

T (x) = Ax for all x ∈ Rn .

4 / 15
Linear transformations from Rn to Rm
• Let A ∈ Rm×n . Define TA : Rn −→ Rm by TA (x) := Ax for all x ∈ Rn . Then TA is
linear. Indeed, TA (αx + βy) = A(αx + βy) = αAx + βAy = αTA (x) + βTA (y).

• If T : Rn −→ Rm is a linear transformation, then there exists A ∈ Rm×n such that

h i
T (x) = Ax for all x ∈ Rn . The matrix A is given by A = T (e1 ) · · · T (en ) .

4 / 15
Linear transformations from Rn to Rm
• Let A ∈ Rm×n . Define TA : Rn −→ Rm by TA (x) := Ax for all x ∈ Rn . Then TA is
linear. Indeed, TA (αx + βy) = A(αx + βy) = αAx + βAy = αTA (x) + βTA (y).

• If T : Rn −→ Rm is a linear transformation, then there exists A ∈ Rm×n such that

h i
T (x) = Ax for all x ∈ Rn . The matrix A is given by A = T (e1 ) · · · T (en ) .
h i>
Proof: For all x := x1 · · · xn ∈ Rn , we have
h i
T (x) = T (x1 e1 + · · · + xn en ) = x1 T (e1 ) + · · · + xn T (en ) = T (e1 ) · · · T (en ) x. 

4 / 15
Linear transformations from Rn to Rm
• Let A ∈ Rm×n . Define TA : Rn −→ Rm by TA (x) := Ax for all x ∈ Rn . Then TA is
linear. Indeed, TA (αx + βy) = A(αx + βy) = αAx + βAy = αTA (x) + βTA (y).

• If T : Rn −→ Rm is a linear transformation, then there exists A ∈ Rm×n such that

h i
T (x) = Ax for all x ∈ Rn . The matrix A is given by A = T (e1 ) · · · T (en ) .
h i>
Proof: For all x := x1 · · · xn ∈ Rn , we have
h i
T (x) = T (x1 e1 + · · · + xn en ) = x1 T (e1 ) + · · · + xn T (en ) = T (e1 ) · · · T (en ) x. 

Remark: An m × n matrix A can be identified with the linear transformation TA and

hence A can be treated as a linear transformation from Rn to Rm .

4 / 15
Linear transformations from Rn to Rm
• Let A ∈ Rm×n . Define TA : Rn −→ Rm by TA (x) := Ax for all x ∈ Rn . Then TA is
linear. Indeed, TA (αx + βy) = A(αx + βy) = αAx + βAy = αTA (x) + βTA (y).

• If T : Rn −→ Rm is a linear transformation, then there exists A ∈ Rm×n such that

h i
T (x) = Ax for all x ∈ Rn . The matrix A is given by A = T (e1 ) · · · T (en ) .
h i>
Proof: For all x := x1 · · · xn ∈ Rn , we have
h i
T (x) = T (x1 e1 + · · · + xn en ) = x1 T (e1 ) + · · · + xn T (en ) = T (e1 ) · · · T (en ) x. 

Remark: An m × n matrix A can be identified with the linear transformation TA and

hence A can be treated as a linear transformation
" #! " from R#n"to#Rm .
x cos θ − sin θ x
• Define T : R2 −→ R2 by T := . Then
y sin θ cos θ y

4 / 15
Linear transformations from Rn to Rm
• Let A ∈ Rm×n . Define TA : Rn −→ Rm by TA (x) := Ax for all x ∈ Rn . Then TA is
linear. Indeed, TA (αx + βy) = A(αx + βy) = αAx + βAy = αTA (x) + βTA (y).

• If T : Rn −→ Rm is a linear transformation, then there exists A ∈ Rm×n such that

h i
T (x) = Ax for all x ∈ Rn . The matrix A is given by A = T (e1 ) · · · T (en ) .
h i>
Proof: For all x := x1 · · · xn ∈ Rn , we have
h i
T (x) = T (x1 e1 + · · · + xn en ) = x1 T (e1 ) + · · · + xn T (en ) = T (e1 ) · · · T (en ) x. 

Remark: An m × n matrix A can be identified with the linear transformation TA and

hence A can be treated as a linear transformation
" #! " from R#n"to#Rm .
x cos θ − sin θ x
• Define T : R2 −→ R2 by T := . Then T is a linear
y sin θ cos θ y
transformation that rotates vectors in R2 by an angle θ in the anti-clockwise direction.
4 / 15
Plane rotation "in #R2 " #" # " #
x cos θ − sin θ x x cos θ − y sin θ
We have G (θ) = = .
y sin θ cos θ y x sin θ + y cos θ

5 / 15
Plane rotation "in #R2 " #" # " #
x cos θ − sin θ x x cos θ − y sin θ
We have G (θ) = = .
y sin θ cos θ y x sin θ + y cos θ
e2

" #
cos θ
= G (θ)e1
sin θ

θ
e1
Plane rotation "in #R2 " #" # " #
x cos θ − sin θ x x cos θ − y sin θ
We have G (θ) = = .
y sin θ cos θ y x sin θ + y cos θ
e2

" #
cos θ
= G (θ)e1
sin θ

θ
e1
Plane rotation "in #R2 " #" # " #
x cos θ − sin θ x x cos θ − y sin θ
We have G (θ) = = .
y sin θ cos θ y x sin θ + y cos θ

" # e2
− sin θ
G (θ)e2 =
cos θ
" #
cos θ
= G (θ)e1
θ sin θ

θ
e1
Plane rotation "in #R2 " #" # " #
x cos θ − sin θ x x cos θ − y sin θ
We have G (θ) = = .
y sin θ cos θ y x sin θ + y cos θ

" # e2
− sin θ
G (θ)e2 =
cos θ
" #
cos θ
= G (θ)e1
θ sin θ

θ
e1
Plane rotation "in #R2 " #" # " #
x cos θ − sin θ x x cos θ − y sin θ
We have G (θ) = = .
y sin θ cos θ y x sin θ + y cos θ

" # e2
− sin θ
G (θ)e2 =
cos θ
" #
cos θ
= G (θ)e1
θ sin θ

θ
e1

The rotation G (θ). Note that G (θ)e2 = G (π/2 + θ)e1 .

5 / 15
Examples
df (x)
Define T : C 1 [0, 1] → C [0, 1] by (Tf )(x) := dx . Then T is a linear
transformation.

6 / 15
Examples
df (x)
Define T : C 1 [0, 1] → C [0, 1] by (Tf )(x) := dx . Then T is a linear
transformation.
Rx
Define T : C [0, 1] → C [0, 1] by (Tf )(x) := 0 f (t)dt. Then T is a linear
transformation.

6 / 15
Examples
df (x)
Define T : C 1 [0, 1] → C [0, 1] by (Tf )(x) := dx . Then T is a linear
transformation.
Rx
Define T : C [0, 1] → C [0, 1] by (Tf )(x) := 0 f (t)dt. Then T is a linear
transformation.
Define T : R[x] → R by T (p(x)) := p(0). Then T is a linear transformation.

6 / 15
Examples
df (x)
Define T : C 1 [0, 1] → C [0, 1] by (Tf )(x) := dx . Then T is a linear
transformation.
Rx
Define T : C [0, 1] → C [0, 1] by (Tf )(x) := 0 f (t)dt. Then T is a linear
transformation.
Define T : R[x] → R by T (p(x)) := p(0). Then T is a linear transformation.
Define T : Rn×n −→ Rn×n by T (A) := 2A − A> . Then T is a linear
transformation.

6 / 15
Examples
df (x)
Define T : C 1 [0, 1] → C [0, 1] by (Tf )(x) := dx . Then T is a linear
transformation.
Rx
Define T : C [0, 1] → C [0, 1] by (Tf )(x) := 0 f (t)dt. Then T is a linear
transformation.
Define T : R[x] → R by T (p(x)) := p(0). Then T is a linear transformation.
Define T : Rn×n −→ Rn×n by T (A) := 2A − A> . Then T is a linear
transformation.
Let A ∈ Rm×m and B ∈ Rn×n . Define T : Rm×n −→ Rm×n by
T (X ) := AX − XB. Then T is a linear transformation.

6 / 15
Examples
df (x)
Define T : C 1 [0, 1] → C [0, 1] by (Tf )(x) := dx . Then T is a linear
transformation.
Rx
Define T : C [0, 1] → C [0, 1] by (Tf )(x) := 0 f (t)dt. Then T is a linear
transformation.
Define T : R[x] → R by T (p(x)) := p(0). Then T is a linear transformation.
Define T : Rn×n −→ Rn×n by T (A) := 2A − A> . Then T is a linear
transformation.
Let A ∈ Rm×m and B ∈ Rn×n . Define T : Rm×n −→ Rm×n by
T (X ) := AX − XB. Then T is a linear transformation.
" #
p(1) p(2)
Define T : Pn −→ R2×2 by T (p(x)) := . Then T is a linear
p(3) p(4)
transformation.
6 / 15
Basic properties of LT

Fact: Let T : V → W be an LT. Then

1 T (0) = 0;

7 / 15
Basic properties of LT

Fact: Let T : V → W be an LT. Then

1 T (0) = 0; [T (0) = T (0 + 0) = T (0) + T (0) ⇒ 0 = T (0).]

7 / 15
Basic properties of LT

Fact: Let T : V → W be an LT. Then

1 T (0) = 0; [T (0) = T (0 + 0) = T (0) + T (0) ⇒ 0 = T (0).]
2 T (−v) = −T (v) for all v ∈ V;

7 / 15
Basic properties of LT

Fact: Let T : V → W be an LT. Then

1 T (0) = 0; [T (0) = T (0 + 0) = T (0) + T (0) ⇒ 0 = T (0).]
2 T (−v) = −T (v) for all v ∈ V; [T (−v) + T v = 0.]

7 / 15
Basic properties of LT

Fact: Let T : V → W be an LT. Then

1 T (0) = 0; [T (0) = T (0 + 0) = T (0) + T (0) ⇒ 0 = T (0).]
2 T (−v) = −T (v) for all v ∈ V; [T (−v) + T v = 0.]
3 T (u − v) = T (u) − T (v) for all u, v ∈ V.

7 / 15
Basic properties of LT

Fact: Let T : V → W be an LT. Then

1 T (0) = 0; [T (0) = T (0 + 0) = T (0) + T (0) ⇒ 0 = T (0).]
2 T (−v) = −T (v) for all v ∈ V; [T (−v) + T v = 0.]
3 T (u − v) = T (u) − T (v) for all u, v ∈ V.
2 2
Question:
" #! Does" #there exist
" #!a linear
" #transformation
" #! " #→ R such that
T :R
0 2 1 3 1 3
T = , T = and T = ?
1 3 0 2 1 3

7 / 15
Basic properties of LT

Fact: Let T : V → W be an LT. Then

1 T (0) = 0; [T (0) = T (0 + 0) = T (0) + T (0) ⇒ 0 = T (0).]
2 T (−v) = −T (v) for all v ∈ V; [T (−v) + T v = 0.]
3 T (u − v) = T (u) − T (v) for all u, v ∈ V.
2 2
Question:
" #! Does" #there exist
" #!a linear
" #transformation
" #! T :R " #→ R such that
0 2 1 3 1 3
T = , T = and T = ?
1 3 0 2 1 3
" # " #! " #
3 1 5
Answer: No, because =T = T (e1 ) + T (e2 ) = is not possible.
3 1 5

7 / 15
Action of an LT
Theorem: Let {v1 , . . . , vn } be a basis of V. Let u1 , . . . , un be arbitrary vectors in W.
Then there is a unique LT T : V −→ W such that T (vi ) = ui for i = 1 : n.

8 / 15
Action of an LT
Theorem: Let {v1 , . . . , vn } be a basis of V. Let u1 , . . . , un be arbitrary vectors in W.
Then there is a unique LT T : V −→ W such that T (vi ) = ui for i = 1 : n.

Proof: Every v ∈ V can be written uniquely as v = α1 v1 + · · · + αn vn .

8 / 15
Action of an LT
Theorem: Let {v1 , . . . , vn } be a basis of V. Let u1 , . . . , un be arbitrary vectors in W.
Then there is a unique LT T : V −→ W such that T (vi ) = ui for i = 1 : n.

Proof: Every v ∈ V can be written uniquely as v = α1 v1 + · · · + αn vn . Define

T : V −→ W by T (α1 v1 + · · · + αn vn ) := α1 u1 + · · · + αn un ∈ W. Then T is the
required LT. 

8 / 15
Action of an LT
Theorem: Let {v1 , . . . , vn } be a basis of V. Let u1 , . . . , un be arbitrary vectors in W.
Then there is a unique LT T : V −→ W such that T (vi ) = ui for i = 1 : n.

Proof: Every v ∈ V can be written uniquely as v = α1 v1 + · · · + αn vn . Define

T : V −→ W by T (α1 v1 + · · · + αn vn ) := α1 u1 + · · · + αn un ∈ W. Then T is the
required LT. 

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

the domain space.

8 / 15
Action of an LT
Theorem: Let {v1 , . . . , vn } be a basis of V. Let u1 , . . . , un be arbitrary vectors in W.
Then there is a unique LT T : V −→ W such that T (vi ) = ui for i = 1 : n.

Proof: Every v ∈ V can be written uniquely as v = α1 v1 + · · · + αn vn . Define

T : V −→ W by T (α1 v1 + · · · + αn vn ) := α1 u1 + · · · + αn un ∈ W. Then T is the
required LT. 

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

the domain space.

Example: Let T : R2 → P2 be linear and T (e1 ) = 2 − 3x + x 2 and T (e2 ) = 1 − x 2 .

8 / 15
Action of an LT
Theorem: Let {v1 , . . . , vn } be a basis of V. Let u1 , . . . , un be arbitrary vectors in W.
Then there is a unique LT T : V −→ W such that T (vi ) = ui for i = 1 : n.

Proof: Every v ∈ V can be written uniquely as v = α1 v1 + · · · + αn vn . Define

T : V −→ W by T (α1 v1 + · · · + αn vn ) := α1 u1 + · · · + αn un ∈ W. Then T is the
required LT. 

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

the domain space.

Example: Let T : R2 → P2 be linear and T (e1 ) = 2 − 3x + x 2 and T (e2 ) = 1 − x 2 .

Determine T ([2, 3]> ).

8 / 15
Action of an LT
Theorem: Let {v1 , . . . , vn } be a basis of V. Let u1 , . . . , un be arbitrary vectors in W.
Then there is a unique LT T : V −→ W such that T (vi ) = ui for i = 1 : n.

Proof: Every v ∈ V can be written uniquely as v = α1 v1 + · · · + αn vn . Define

T : V −→ W by T (α1 v1 + · · · + αn vn ) := α1 u1 + · · · + αn un ∈ W. Then T is the
required LT. 

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

the domain space.

Example: Let T : R2 → P2 be linear and T (e1 ) = 2 − 3x + x 2 and T (e2 ) = 1 − x 2 .

Determine >
" #! T ([2, 3] ). We have
2
T = T (2e1 +3e2 ) = 2T (e1 )+3T (e2 ) = 2(2−3x +x 2 )+3(1−x 2 ) = 7−6x −x 2 .
3
8 / 15
Kernel and range space
Let T : V −→ W be a linear transformation. Then the kernel (null space) N(T ) and
the range space R(T ) of T are defined by

N(T ) := {v ∈ V : T v = 0} ⊆ V and R(T ) := {T v : v ∈ V} ⊆ W.

9 / 15
Kernel and range space
Let T : V −→ W be a linear transformation. Then the kernel (null space) N(T ) and
the range space R(T ) of T are defined by

N(T ) := {v ∈ V : T v = 0} ⊆ V and R(T ) := {T v : v ∈ V} ⊆ W.

Remark: The kernel of T is also denoted by ker(T ).

9 / 15
Kernel and range space
Let T : V −→ W be a linear transformation. Then the kernel (null space) N(T ) and
the range space R(T ) of T are defined by

N(T ) := {v ∈ V : T v = 0} ⊆ V and R(T ) := {T v : v ∈ V} ⊆ W.

Remark: The kernel of T is also denoted by ker(T ). If T : Rn −→ Rm is given by

T (x) := Ax, where A ∈ Rm×n , then N(T ) = null(A) and R(T ) = Col(A).

9 / 15
Kernel and range space
Let T : V −→ W be a linear transformation. Then the kernel (null space) N(T ) and
the range space R(T ) of T are defined by

N(T ) := {v ∈ V : T v = 0} ⊆ V and R(T ) := {T v : v ∈ V} ⊆ W.

Remark: The kernel of T is also denoted by ker(T ). If T : Rn −→ Rm is given by

T (x) := Ax, where A ∈ Rm×n , then N(T ) = null(A) and R(T ) = Col(A).

Theorem: We have N(T ) 4 V and R(T ) 4 W.

9 / 15
Kernel and range space
Let T : V −→ W be a linear transformation. Then the kernel (null space) N(T ) and
the range space R(T ) of T are defined by

N(T ) := {v ∈ V : T v = 0} ⊆ V and R(T ) := {T v : v ∈ V} ⊆ W.

Remark: The kernel of T is also denoted by ker(T ). If T : Rn −→ Rm is given by

T (x) := Ax, where A ∈ Rm×n , then N(T ) = null(A) and R(T ) = Col(A).

Theorem: We have N(T ) 4 V and R(T ) 4 W. If B := {v1 , . . . , vm } spans V, then

T (B) = {T (v1 ), . . . , T (vm )} spans R(T ).

9 / 15
Kernel and range space
Let T : V −→ W be a linear transformation. Then the kernel (null space) N(T ) and
the range space R(T ) of T are defined by

N(T ) := {v ∈ V : T v = 0} ⊆ V and R(T ) := {T v : v ∈ V} ⊆ W.

Remark: The kernel of T is also denoted by ker(T ). If T : Rn −→ Rm is given by

T (x) := Ax, where A ∈ Rm×n , then N(T ) = null(A) and R(T ) = Col(A).

Theorem: We have N(T ) 4 V and R(T ) 4 W. If B := {v1 , . . . , vm } spans V, then

T (B) = {T (v1 ), . . . , T (vm )} spans R(T ).

Proof: Let w ∈ R(T ). Then T (v) = w for some v ∈ V. Now

v = c1 v1 + · · · + cm vm =⇒ w = T (v) = c1 T (v1 ) + · · · + cm T (vm ). 

9 / 15
Rank-nullity theorem

Definition: Let T : V → W be a linear transformation. Define rank(T ) := dim(R(T ))

and nullity(T ) := dim(N(T )).

10 / 15
Rank-nullity theorem

Definition: Let T : V → W be a linear transformation. Define rank(T ) := dim(R(T ))

and nullity(T ) := dim(N(T )).

The Rank-Nullity Theorem: Let V be finite dimensional and T : V −→ W be a linear

transformation. Then rank(T ) + nullity(T ) = dim(V).

10 / 15
Rank-nullity theorem

Definition: Let T : V → W be a linear transformation. Define rank(T ) := dim(R(T ))

and nullity(T ) := dim(N(T )).

The Rank-Nullity Theorem: Let V be finite dimensional and T : V −→ W be a linear

transformation. Then rank(T ) + nullity(T ) = dim(V).
Proof: Suppose that dim(V) = n. Let B = {v1 , . . . , vp } be a basis of N(T ).

10 / 15
Rank-nullity theorem

Definition: Let T : V → W be a linear transformation. Define rank(T ) := dim(R(T ))

and nullity(T ) := dim(N(T )).

The Rank-Nullity Theorem: Let V be finite dimensional and T : V −→ W be a linear

transformation. Then rank(T ) + nullity(T ) = dim(V).
Proof: Suppose that dim(V) = n. Let B = {v1 , . . . , vp } be a basis of N(T ). Extend B
to a basis B ∪ {vp+1 , . . . , vn } of V.

10 / 15
Rank-nullity theorem

Definition: Let T : V → W be a linear transformation. Define rank(T ) := dim(R(T ))

and nullity(T ) := dim(N(T )).

The Rank-Nullity Theorem: Let V be finite dimensional and T : V −→ W be a linear

transformation. Then rank(T ) + nullity(T ) = dim(V).
Proof: Suppose that dim(V) = n. Let B = {v1 , . . . , vp } be a basis of N(T ). Extend B
to a basis B ∪ {vp+1 , . . . , vn } of V. Then {T (vp+1 ), . . . , T (vn )} spans R(T ).

10 / 15
Rank-nullity theorem

Definition: Let T : V → W be a linear transformation. Define rank(T ) := dim(R(T ))

and nullity(T ) := dim(N(T )).

The Rank-Nullity Theorem: Let V be finite dimensional and T : V −→ W be a linear

transformation. Then rank(T ) + nullity(T ) = dim(V).
Proof: Suppose that dim(V) = n. Let B = {v1 , . . . , vp } be a basis of N(T ). Extend B
to a basis B ∪ {vp+1 , . . . , vn } of V. Then {T (vp+1 ), . . . , T (vn )} spans R(T ). In fact,
{T (vp+1 ), . . . , T (vn )} is linearly independent.

10 / 15
Rank-nullity theorem

Definition: Let T : V → W be a linear transformation. Define rank(T ) := dim(R(T ))

and nullity(T ) := dim(N(T )).

The Rank-Nullity Theorem: Let V be finite dimensional and T : V −→ W be a linear

transformation. Then rank(T ) + nullity(T ) = dim(V).
Proof: Suppose that dim(V) = n. Let B = {v1 , . . . , vp } be a basis of N(T ). Extend B
to a basis B ∪ {vp+1 , . . . , vn } of V. Then {T (vp+1 ), . . . , T (vn )} spans R(T ). In fact,
{T (vp+1 ), . . . , T (vn )} is linearly independent. Indeed, let cp+1 , . . . , cn be such that
cp+1 T (vp+1 ) + · · · + cn T (vn ) = 0 =⇒ T (cp+1 vp+1 + · · · + cn vn ) = 0.

10 / 15
Rank-nullity theorem

Definition: Let T : V → W be a linear transformation. Define rank(T ) := dim(R(T ))

and nullity(T ) := dim(N(T )).

The Rank-Nullity Theorem: Let V be finite dimensional and T : V −→ W be a linear

transformation. Then rank(T ) + nullity(T ) = dim(V).
Proof: Suppose that dim(V) = n. Let B = {v1 , . . . , vp } be a basis of N(T ). Extend B
to a basis B ∪ {vp+1 , . . . , vn } of V. Then {T (vp+1 ), . . . , T (vn )} spans R(T ). In fact,
{T (vp+1 ), . . . , T (vn )} is linearly independent. Indeed, let cp+1 , . . . , cn be such that
cp+1 T (vp+1 ) + · · · + cn T (vn ) = 0 =⇒ T (cp+1 vp+1 + · · · + cn vn ) = 0. Consequently,
cp+1 vp+1 + · · · + cn vn ∈ N(T ) =⇒ c1 v1 + · · · + cp vp + cp+1 vp+1 + · · · + cn vn = 0

10 / 15
Rank-nullity theorem

Definition: Let T : V → W be a linear transformation. Define rank(T ) := dim(R(T ))

and nullity(T ) := dim(N(T )).

The Rank-Nullity Theorem: Let V be finite dimensional and T : V −→ W be a linear

transformation. Then rank(T ) + nullity(T ) = dim(V).
Proof: Suppose that dim(V) = n. Let B = {v1 , . . . , vp } be a basis of N(T ). Extend B
to a basis B ∪ {vp+1 , . . . , vn } of V. Then {T (vp+1 ), . . . , T (vn )} spans R(T ). In fact,
{T (vp+1 ), . . . , T (vn )} is linearly independent. Indeed, let cp+1 , . . . , cn be such that
cp+1 T (vp+1 ) + · · · + cn T (vn ) = 0 =⇒ T (cp+1 vp+1 + · · · + cn vn ) = 0. Consequently,
cp+1 vp+1 + · · · + cn vn ∈ N(T ) =⇒ c1 v1 + · · · + cp vp + cp+1 vp+1 + · · · + cn vn = 0
which gives c1 = · · · = cn = 0. Hence rank(T ) = n − p = dim(V) − nullity(T ). 

10 / 15
Example
Let D : P3 → P3 be defined by D(p(x)) = p 0 (x).

11 / 15
Example
Let D : P3 → P3 be defined by D(p(x)) = p 0 (x). Then

N(D) = {a + bx + cx 2 + dx 3 : D(a + bx + cx 2 + dx 3 ) = 0}

11 / 15
Example
Let D : P3 → P3 be defined by D(p(x)) = p 0 (x). Then

N(D) = {a + bx + cx 2 + dx 3 : D(a + bx + cx 2 + dx 3 ) = 0}
= {a + bx + cx 2 + dx 3 : bx + 2cx + 3dx 2 = 0}

11 / 15
Example
Let D : P3 → P3 be defined by D(p(x)) = p 0 (x). Then

N(D) = {a + bx + cx 2 + dx 3 : D(a + bx + cx 2 + dx 3 ) = 0}
= {a + bx + cx 2 + dx 3 : bx + 2cx + 3dx 2 = 0}
= {a + bx + cx 2 + dx 3 : b = c = d = 0}

11 / 15
Example
Let D : P3 → P3 be defined by D(p(x)) = p 0 (x). Then

N(D) = {a + bx + cx 2 + dx 3 : D(a + bx + cx 2 + dx 3 ) = 0}
= {a + bx + cx 2 + dx 3 : bx + 2cx + 3dx 2 = 0}
= {a + bx + cx 2 + dx 3 : b = c = d = 0} = {a : a ∈ R}.

11 / 15
Example
Let D : P3 → P3 be defined by D(p(x)) = p 0 (x). Then

N(D) = {a + bx + cx 2 + dx 3 : D(a + bx + cx 2 + dx 3 ) = 0}
= {a + bx + cx 2 + dx 3 : bx + 2cx + 3dx 2 = 0}
= {a + bx + cx 2 + dx 3 : b = c = d = 0} = {a : a ∈ R}.

Thus N(D) consists of constant polynomials. Hence nullity(D) = 1 and

rank(D) = 4 − 1 = 3.

11 / 15
Example
Let D : P3 → P3 be defined by D(p(x)) = p 0 (x). Then

N(D) = {a + bx + cx 2 + dx 3 : D(a + bx + cx 2 + dx 3 ) = 0}
= {a + bx + cx 2 + dx 3 : bx + 2cx + 3dx 2 = 0}
= {a + bx + cx 2 + dx 3 : b = c = d = 0} = {a : a ∈ R}.

Thus N(D) consists of constant polynomials. Hence nullity(D) = 1 and

rank(D) = 4 − 1 = 3. Next, we have

R(D) = {D(a + bx + cx 2 + dx 3 ) : a, b, c, d ∈ R}

11 / 15
Example
Let D : P3 → P3 be defined by D(p(x)) = p 0 (x). Then

N(D) = {a + bx + cx 2 + dx 3 : D(a + bx + cx 2 + dx 3 ) = 0}
= {a + bx + cx 2 + dx 3 : bx + 2cx + 3dx 2 = 0}
= {a + bx + cx 2 + dx 3 : b = c = d = 0} = {a : a ∈ R}.

Thus N(D) consists of constant polynomials. Hence nullity(D) = 1 and

rank(D) = 4 − 1 = 3. Next, we have

R(D) = {D(a + bx + cx 2 + dx 3 ) : a, b, c, d ∈ R}
= {b + 2cx + 3dx 2 : b, c, d ∈ R}

11 / 15
Example
Let D : P3 → P3 be defined by D(p(x)) = p 0 (x). Then

N(D) = {a + bx + cx 2 + dx 3 : D(a + bx + cx 2 + dx 3 ) = 0}
= {a + bx + cx 2 + dx 3 : bx + 2cx + 3dx 2 = 0}
= {a + bx + cx 2 + dx 3 : b = c = d = 0} = {a : a ∈ R}.

Thus N(D) consists of constant polynomials. Hence nullity(D) = 1 and

rank(D) = 4 − 1 = 3. Next, we have

R(D) = {D(a + bx + cx 2 + dx 3 ) : a, b, c, d ∈ R}
= {b + 2cx + 3dx 2 : b, c, d ∈ R} = P2 .

11 / 15
Example
Let D : P3 → P3 be defined by D(p(x)) = p 0 (x). Then

N(D) = {a + bx + cx 2 + dx 3 : D(a + bx + cx 2 + dx 3 ) = 0}
= {a + bx + cx 2 + dx 3 : bx + 2cx + 3dx 2 = 0}
= {a + bx + cx 2 + dx 3 : b = c = d = 0} = {a : a ∈ R}.

Thus N(D) consists of constant polynomials. Hence nullity(D) = 1 and

rank(D) = 4 − 1 = 3. Next, we have

R(D) = {D(a + bx + cx 2 + dx 3 ) : a, b, c, d ∈ R}
= {b + 2cx + 3dx 2 : b, c, d ∈ R} = P2 .

This shows that rank(D) = dim(P2 ) = 3.

11 / 15
One-one and onto maps
Definition: A linear transformation T : V −→ W is called one-one if T maps distinct
vectors in V to distinct vectors in W.

12 / 15
One-one and onto maps
Definition: A linear transformation T : V −→ W is called one-one if T maps distinct
vectors in V to distinct vectors in W. If R(T ) = W then T is called onto.

12 / 15
One-one and onto maps
Definition: A linear transformation T : V −→ W is called one-one if T maps distinct
vectors in V to distinct vectors in W. If R(T ) = W then T is called onto.
Theorem: Let T : V −→ W be a linear transformation. Then
T is one-one ⇐⇒ N(T ) = {0}.

12 / 15
One-one and onto maps
Definition: A linear transformation T : V −→ W is called one-one if T maps distinct
vectors in V to distinct vectors in W. If R(T ) = W then T is called onto.
Theorem: Let T : V −→ W be a linear transformation. Then
T is one-one ⇐⇒ N(T ) = {0}.

12 / 15
One-one and onto maps
Definition: A linear transformation T : V −→ W is called one-one if T maps distinct
vectors in V to distinct vectors in W. If R(T ) = W then T is called onto.
Theorem: Let T : V −→ W be a linear transformation. Then
T is one-one ⇐⇒ N(T ) = {0}.
If dim(V) = dim(W) = n, then T is onto ⇐⇒ T is one-one ⇐⇒ rank(T ) = n.

12 / 15
One-one and onto maps
Definition: A linear transformation T : V −→ W is called one-one if T maps distinct
vectors in V to distinct vectors in W. If R(T ) = W then T is called onto.
Theorem: Let T : V −→ W be a linear transformation. Then
T is one-one ⇐⇒ N(T ) = {0}.
If dim(V) = dim(W) = n, then T is onto ⇐⇒ T is one-one ⇐⇒ rank(T ) = n.
Proof: Note that T (u) = T (v) ⇐⇒ T (u − v) = 0 ⇐⇒ u − v ∈ N(T ).

12 / 15
One-one and onto maps
Definition: A linear transformation T : V −→ W is called one-one if T maps distinct
vectors in V to distinct vectors in W. If R(T ) = W then T is called onto.
Theorem: Let T : V −→ W be a linear transformation. Then
T is one-one ⇐⇒ N(T ) = {0}.
If dim(V) = dim(W) = n, then T is onto ⇐⇒ T is one-one ⇐⇒ rank(T ) = n.
Proof: Note that T (u) = T (v) ⇐⇒ T (u − v) = 0 ⇐⇒ u − v ∈ N(T ). Hence T is
one-one ⇐⇒ N(T ) = {0}.

12 / 15
One-one and onto maps
Definition: A linear transformation T : V −→ W is called one-one if T maps distinct
vectors in V to distinct vectors in W. If R(T ) = W then T is called onto.
Theorem: Let T : V −→ W be a linear transformation. Then
T is one-one ⇐⇒ N(T ) = {0}.
If dim(V) = dim(W) = n, then T is onto ⇐⇒ T is one-one ⇐⇒ rank(T ) = n.
Proof: Note that T (u) = T (v) ⇐⇒ T (u − v) = 0 ⇐⇒ u − v ∈ N(T ). Hence T is
one-one ⇐⇒ N(T ) = {0}. Next rank(T ) = n ⇐⇒ R(T ) = W ⇐⇒ T is onto. 

12 / 15
One-one and onto maps
Definition: A linear transformation T : V −→ W is called one-one if T maps distinct
vectors in V to distinct vectors in W. If R(T ) = W then T is called onto.
Theorem: Let T : V −→ W be a linear transformation. Then
T is one-one ⇐⇒ N(T ) = {0}.
If dim(V) = dim(W) = n, then T is onto ⇐⇒ T is one-one ⇐⇒ rank(T ) = n.
Proof: Note that T (u) = T (v) ⇐⇒ T (u − v) = 0 ⇐⇒ u − v ∈ N(T ). Hence T is
one-one ⇐⇒ N(T ) = {0}. Next rank(T ) = n ⇐⇒ R(T ) = W ⇐⇒ T is onto. 

Theorem: Let T : V −→ W be a one-one linear transformation. If S := {v1 , . . . , vp } is

a linearly independent set in V, then T (S) = {T (v1 ), . . . , T (vp )} ⊆ W is a linearly
independent set in W.

12 / 15
One-one and onto maps
Definition: A linear transformation T : V −→ W is called one-one if T maps distinct
vectors in V to distinct vectors in W. If R(T ) = W then T is called onto.
Theorem: Let T : V −→ W be a linear transformation. Then
T is one-one ⇐⇒ N(T ) = {0}.
If dim(V) = dim(W) = n, then T is onto ⇐⇒ T is one-one ⇐⇒ rank(T ) = n.
Proof: Note that T (u) = T (v) ⇐⇒ T (u − v) = 0 ⇐⇒ u − v ∈ N(T ). Hence T is
one-one ⇐⇒ N(T ) = {0}. Next rank(T ) = n ⇐⇒ R(T ) = W ⇐⇒ T is onto. 

Theorem: Let T : V −→ W be a one-one linear transformation. If S := {v1 , . . . , vp } is

a linearly independent set in V, then T (S) = {T (v1 ), . . . , T (vp )} ⊆ W is a linearly
independent set in W.
Proof: Let c1 , . . . , cn be scalars such that c1 T (v1 ) + · · · + cn T (vn ) = 0. Then
T (c1 v1 + · · · + cn vn ) = 0 =⇒ c1 v1 + · · · + cn vn = 0 =⇒ c1 = · · · = cn = 0. 
12 / 15
Composition of linear transformations

Composition: Let T : U → V and S : V → W be LTs.

13 / 15
Composition of linear transformations

Composition: Let T : U → V and S : V → W be LTs. Then the composition

S ◦ T : U → W defined by (S ◦ T )u = S(T u) is also a linear transformation.

13 / 15
Composition of linear transformations

Composition: Let T : U → V and S : V → W be LTs. Then the composition

S ◦ T : U → W defined by (S ◦ T )u = S(T u) is also a linear transformation. The
composition S ◦ T is simply written as ST .

13 / 15
Composition of linear transformations

Composition: Let T : U → V and S : V → W be LTs. Then the composition

S ◦ T : U → W defined by (S ◦ T )u = S(T u) is also a linear transformation. The
composition S ◦ T is simply written as ST .
Proof: (S ◦ T )(αu + βv) = S(αT u + βT v) = α(S ◦T )u + β(S ◦ T )v. 

13 / 15
Invertible linear transformations
Definition: A linear transformation T : V −→ W is invertible if there is a linear
transformation S : W −→ V such that S ◦ T = IV and T ◦ S = IW ,

14 / 15
Invertible linear transformations
Definition: A linear transformation T : V −→ W is invertible if there is a linear
transformation S : W −→ V such that S ◦ T = IV and T ◦ S = IW , where IV and IW
are identity transformations on V and W, respectively.

14 / 15
Invertible linear transformations
Definition: A linear transformation T : V −→ W is invertible if there is a linear
transformation S : W −→ V such that S ◦ T = IV and T ◦ S = IW , where IV and IW
are identity transformations on V and W, respectively. The transformation S is called
an inverse of T .

14 / 15
Invertible linear transformations
Definition: A linear transformation T : V −→ W is invertible if there is a linear
transformation S : W −→ V such that S ◦ T = IV and T ◦ S = IW , where IV and IW
are identity transformations on V and W, respectively. The transformation S is called
an inverse of T .

Fact: If a linear transformation T : V −→ W is invertible, then its inverse is unique.

The inverse of T is denoted by T −1 .

14 / 15
Invertible linear transformations
Definition: A linear transformation T : V −→ W is invertible if there is a linear
transformation S : W −→ V such that S ◦ T = IV and T ◦ S = IW , where IV and IW
are identity transformations on V and W, respectively. The transformation S is called
an inverse of T .

Fact: If a linear transformation T : V −→ W is invertible, then its inverse is unique.

The inverse of T is denoted by T −1 .

Theorem: Let T : V −→ W be a linear transformation. Suppose that

dim(V) = dim(W) = n. Then T is invertible ⇐⇒ T is one-one ⇐⇒ T is onto
⇐⇒ rank(T ) = n.

14 / 15
Invertible linear transformations
Definition: A linear transformation T : V −→ W is invertible if there is a linear
transformation S : W −→ V such that S ◦ T = IV and T ◦ S = IW , where IV and IW
are identity transformations on V and W, respectively. The transformation S is called
an inverse of T .

Fact: If a linear transformation T : V −→ W is invertible, then its inverse is unique.

The inverse of T is denoted by T −1 .

Theorem: Let T : V −→ W be a linear transformation. Suppose that

dim(V) = dim(W) = n. Then T is invertible ⇐⇒ T is one-one ⇐⇒ T is onto
⇐⇒ rank(T ) = n.

Proof: Follows from rank-nullity theorem. 

14 / 15
Example " #!
a
• Define T : R2 −→ P1 by T := a + (a + b)x. Then T is invertible and
b
" #
c
T −1 (c + dx) = .
d −c

15 / 15
Example " #!
a
• Define T : R2 −→ P1 by T := a + (a + b)x. Then T is invertible and
b
" # " #! " #
c a a
T −1 (c + dx) = . Indeed, (T −1 ◦ T ) = T −1 (a + (a + b)x) = .
d −c b b

15 / 15
Example " #!
a
• Define T : R2 −→ P1 by T := a + (a + b)x. Then T is invertible and
b
" # " #! " #
c a a
T −1 (c + dx) = . Indeed, (T −1 ◦ T ) = T −1 (a + (a + b)x) = .
d −c b b
" #!
a
Also, (T ◦ T −1 )(a + bx) = T = a + (b − a + a)x = a + bx.
b−a

15 / 15
Example " #!
a
• Define T : R2 −→ P1 by T := a + (a + b)x. Then T is invertible and
b
" # " #! " #
c a a
T −1 (c + dx) = . Indeed, (T −1 ◦ T ) = T −1 (a + (a + b)x) = .
d −c b b
" #!
a
Also, (T ◦ T −1 )(a + bx) = T = a + (b − a + a)x = a + bx.
b−a

• Let A ∈ Rn×n . Define TA : Rn −→ Rn by TA (x) := Ax. Then TA is invertible ⇐⇒ A

is invertible. If A invertible then (TA )−1 = TA−1 .

15 / 15
Example " #!
a
• Define T : R2 −→ P1 by T := a + (a + b)x. Then T is invertible and
b
" # " #! " #
c a a
T −1 (c + dx) = . Indeed, (T −1 ◦ T ) = T −1 (a + (a + b)x) = .
d −c b b
" #!
a
Also, (T ◦ T −1 )(a + bx) = T = a + (b − a + a)x = a + bx.
b−a

• Let A ∈ Rn×n . Define TA : Rn −→ Rn by TA (x) := Ax. Then TA is invertible ⇐⇒ A

is invertible. If A invertible then (TA )−1 = TA−1 .

Proof: Since R(TA ) = Col(A), we have rank(TA ) = rank(A). Hence TA is invertible

⇐⇒ A is invertible.

15 / 15
Example " #!
a
• Define T : R2 −→ P1 by T := a + (a + b)x. Then T is invertible and
b
" # " #! " #
c a a
T −1 (c + dx) = . Indeed, (T −1 ◦ T ) = T −1 (a + (a + b)x) = .
d −c b b
" #!
a
Also, (T ◦ T −1 )(a + bx) = T = a + (b − a + a)x = a + bx.
b−a

• Let A ∈ Rn×n . Define TA : Rn −→ Rn by TA (x) := Ax. Then TA is invertible ⇐⇒ A

is invertible. If A invertible then (TA )−1 = TA−1 .

Proof: Since R(TA ) = Col(A), we have rank(TA ) = rank(A). Hence TA is invertible

⇐⇒ A is invertible. Now (TA−1 ◦ TA )(x) = TA−1 (TA (x)) = TA−1 (Ax) = A−1 Ax = x
shows that TA−1 ◦ TA = In .

15 / 15
Example " #!
a
• Define T : R2 −→ P1 by T := a + (a + b)x. Then T is invertible and
b
" # " #! " #
c a a
T −1 (c + dx) = . Indeed, (T −1 ◦ T ) = T −1 (a + (a + b)x) = .
d −c b b
" #!
a
Also, (T ◦ T −1 )(a + bx) = T = a + (b − a + a)x = a + bx.
b−a

• Let A ∈ Rn×n . Define TA : Rn −→ Rn by TA (x) := Ax. Then TA is invertible ⇐⇒ A

is invertible. If A invertible then (TA )−1 = TA−1 .

Proof: Since R(TA ) = Col(A), we have rank(TA ) = rank(A). Hence TA is invertible

⇐⇒ A is invertible. Now (TA−1 ◦ TA )(x) = TA−1 (TA (x)) = TA−1 (Ax) = A−1 Ax = x
shows that TA−1 ◦ TA = In . Similarly, TA ◦ TA−1 = In . Hence (TA )−1 = TA−1 . 

*** 15 / 15