Delhi School of Economics

EC-002 Introductory Maths

Linear Transformations: Examples
August 31, 2025

LT Examples in R2
1. Stretching Transformation
Transformation: (x, y) → (mx, my), where m is some scalar
Matrix:  
m 0
A=
0 m

Derivation:
We want:      
x mx a b
A = . Think that A =
y my c d

Matching coefficients:

ax + by = mx ⇒ a = m, b = 0

cx + dy = my ⇒ c = 0, d = m

So the matrix is A (as shown above)

2. Rotation by 90° Counterclockwise in R2

Transformation: (x, y) → (−y, x) → general rule.
Notice: but, if you start with vector on X-axis, then?
Matrix:  
0 −1
A=
1 0

Derivation:
We want:    
x −y
A =
y x

Matching coefficients:

ax + by = −y ⇒ a = 0, b = −1

1
 cx + dy = x ⇒ c = 1, d = 0

So the matrix is:    

0 −1 cosθ −sinθ
=
1 0 sinθ cosθ
Note 1: This matrix rotates vectors by 90 degrees counterclockwise. It is derived
from the general rotation matrix for angle θ. For θ = 90◦ , cosθ = 0 , sinθ = 1
Note 2: Rotation by 90° Counterclockwise in R3 . How??.
We need to fix one axis, to say that rotation is happenning around that axis (so
that values in that direction do not change.)
Fixing direction 3, Transformation: (x, y, z) → (−y, x, z)
So the matrix is:
   
π  0 −1 0 cosθ −sinθ 0
Rz = 1 0 0 == sinθ cosθ 0
2
0 0 1 0 0 1

Do you notice what’s going on here? Can you derive these matrices using the
methods shown above?
Also try to derive the rotation matrix for when when we rotate around dimention-1
(i.e. X-axis).

3. Reflection Across the 45◦ Line, i.e. y = x

Transformation: (x, y) → (y, x).
This transformation swaps the x and y coordinates (but signs are preserved)
Matrix:  
0 1
A=
1 0

Derivation:
We want:    
x y
A =
y x

Matching coefficients:
ax + by = y ⇒ a = 0, b = 1
cx + dy = x ⇒ c = 1, d = 0

So the matrix is A as derived above.

◦◦ Comments - Rotation Matrices:

ˆ Rotation Matrices are always square and real.

ˆ Orthogonality: Rows & columns are orthogonal vectors i.e. they are orthogo-
nal (perpendicular) to eacch other. Algebraically: RT = R−1
ˆ Some of entries in every column of R is unity.

2
4. Projection onto the x-axis
Transformation: (x, y) → (x, 0)
Matrix:  
1 0
A=
0 0

Derivation:
We want:    
x x
A =
y 0

Matching coefficients:
ax + by = x ⇒ a = 1, b = 0
cx + dy = 0 ⇒ c = 0, d = 0

So the matrix is:  

1 0
0 0

5. Projection onto the xy-plane in 3D

Transformation: (x, y, z) → (x, y, 0)
Matrix:  
1 0 0
A = 0 1 0
0 0 0

Derivation:
We want:    
x x
A y = y
  
z 0

Matching coefficients:

ax + by + cz = x ⇒ a = 1, b = 0, c = 0

dx + ey + f z = y ⇒ d = 0, e = 1, f = 0
gx + hy + iz = 0 ⇒ g = 0, h = 0, i = 0

So the matrix is as shown above

6. Differentiation of a Polynomial of Degree 3

Let P3 (t) = a0 + a1 t + a2 t2 + a3 t3 .
Basis of P4 : {1, t, t2 , t3 } and Basis of P3 : {1, t, t2 }
Differentiation:
d
P (t) = [0.a0 ] + a1 + 2a2 t + 3a3 t2
dt

3
 Differentiation maps a degree-3 polynomial to a degree-2 polynomial. The matrix
is constructed by applying the derivative to each basis element and expressing the
result in the basis {1, t, t2 }
Derivation: We want:
   
a0 .
 a1   a1 
A3×4  
 a2  = 
2a2 
a3 4×1 3a3 3×1

Notice: for matrix multiplication to be balanced what the dimension of matrix A

be here.
So T : V 4 → W 3 : this case is ”blown down”.
A reverse of this (”blown up”) is also possible, when T : V 3 → W 4 .
Matching coefficients:
Row 1: a1 ⇒ [0, 1, 0, 0]
Row 2: 2a2 ⇒ [0, 0, 2, 0]
Row 3: 3a3 ⇒ [0, 0, 0, 3]

So the matrix is:  

0 1 0 0
0 0 2 0
0 0 0 3

7. Example: Linear Transformation T (u) = −2u

Let T : V → W be a linear transformation where V, W ⊆ R2 . Suppose {v1 , v2 } is a
basis for V . We are given:
T (u) = −2u

Method I: Using Basis Vectors Since {v1 , v2 } is a basis, any v ∈ V can be

written as:
v = x1 v1 + x2 v2

Applying T to both sides:

T (v) = x1 T (v1 ) + x2 T (v2 ) (1)

Given the transformation:

T (v1 ) = −2v1 = −2.v1 + 0.v2 (2)

T (v2 ) = −2v2 = 0.v1 − 2.v2 (3)

So comparing eqns (3) & (4) with eqn (1), the matrix A in the basis {v1 , v2 } satisfies:
 
−2 0
A=
0 −2

4
 Method II: Using Matrix Multiplication Assume the matrix of the linear
transformation T is:  
a11 a12
A=
a21 a22

Then by definition of T , we have: T (v) = Av = −2v.

Therefore, for v1 : (expressing in terms of basis vectors)

T (v1 ) = Av1 = −2v1 = a11 .v1 + a21 .v2

 

a11
=⇒ −2.v1 + 0.v2 = v1 v2
a21
 
a11
−1
=⇒ = v1 v2 .(−2.v1 ) (4)
a21

Similarly, for v2 :
T (v2 ) = Av2 = −2v2 = a12 .v1 + a22 .v2
 

a12
=⇒ 0.v1 + (−2).v2 = v1 v2
a22
 
a12
−1
=⇒ = v1 v2 .(−2.v1 ) (5)
a22

Now from (1) and (2), by the similarity of LHS & RHS, we
Therefore, the matrix is:  
−2 0
A=
0 −2

8. In R2 , compute the matrix of linear transformation for the reflection on the 135°
line.
Sol: The 135° line passes through the origin and makes an angle of 135° with the
positive x-axis.
Method 1: Using the general reflection formula
For reflection across a line making angle θ with the x-axis, the reflection matrix is:
 
cos 2θ sin 2θ
R=
sin 2θ − cos 2θ

With θ = 135, we have 2θ = 270:

cos 270 = 0 (6)

sin 270 = −1 (7)

Therefore:  
0 −1
R=
−1 0

Method 2: Direct computation

Let’s find where the standard basis vectors map:

5
 For e1 = (1, 0): The 135° line has equation y = −x. The reflection of point (1, 0)
across
this line: - The perpendicular from (1, 0) to
 line y = −x meets
the line at
1
2
, − 2 - The reflection is: (1, 0) − 2 (1, 0) − 2 , − 2 = (1, 0) − 2 12 , 12 = (0, −1)
1 1 1

For e2
= (0, 1): - The perpendicularfrom (0, 1) to line
 y = −x meets
the line at
− 21 , 12 - The reflection is: (0, 1) − 2 (0, 1) − − 12 , 12 = (0, 1) − 2 12 , 12 = (−1, 0)
So: - R(e1 ) = R(1, 0) = (0, −1) - R(e2 ) = R(0, 1) = (−1, 0)
Therefore, the matrix is:  
0 −1
R=
−1 0

 135° line
Verification: Let’s check that points on the  are fixed:
 
0 −1 1 1
For point (1, −1) on the line: R(1, −1) = =
−1 0 −1 −1
Finally, the matrix of reflection across the 135° line is:
 
0 −1
R=
−1 0

9. (Strang Sec. 2.3 Q.24.) Find a basis for the plane x − 2y + 3z = 0 in R3 . Then
find a basis for the intersection of that plane with the xy-plane. Then find a basis
for all vectors perpendicular to the plane.
Sol: (a) Basis for the plane x − 2y + 3z = 0 in R3 :
Let y = s, z = t, where s, t ∈ R. Then,

x = 2y − 3z = 2s − 3t

So every vector in the plane can be written as:

     
x 2 −3
y  = s 1 + t  0 
z 0 1

Thus, a basis for the plane is

   
 2 −3 
1 ,  0 
0 1
 

(b) Basis for the intersection of the plane with the xy-plane:
The xy-plane is described by z = 0. Substitute z = 0 into the equation of the plane:

x − 2y = 0 =⇒ x = 2y

So vectors in the intersection take the form:

     
x 2y 2
y  =  y  = y 1
z 0 0

6
 Therefore, a basis is
  
 2 
1
0
 

(c) Basis for all vectors perpendicular to the plane:

The normal vector to the plane x − 2y + 3z = 0 is
 
1
n = −2

3

So a basis for all vectors perpendicular to the plane is

  
 1 
−2
3
 

10. Strang Sec 2.3 Q.10. Find two independent vectors on the plane x+2y−3z−t = 0
in R4 . Then find three independent vectors. Why not four? This plane is the
nullspace of what matrix?

Solution.
Let the general vector (x, y, z, t) ∈ R4 . The equation of the plane can be written:

x + 2y − 3z − t = 0 =⇒ x = −2y + 3z + t

So every vector on the plane is of the form

         
x −2y + 3z + t −2 3 1
y   y  1 0 0
 =  = y  + z  + t 
z   z  0 1 0
t t 0 0 1

Two independent vectors on the plane:

   
−2 3
1 0
 
0 and  
1
0 0

Three independent vectors on the plane:

     
−2 3 1
 1  0 0
 , , 
 0  1 0
0 0 1

7
 Why not four?
Because the plane is defined as the solution set to one linear equation in R4 , that
solution set is a subspace of dimension 4 − 1 = 3. So, its basis has at most three
independent vectors. Sol: This plane is the nullspace of what matrix?
It is the nullspace of the matrix


1 2 −3 −1

since all vectors in the plane satisfy x + 2y − 3z − t = 0.

11. (Treil 1.2.2 (a)) Write down a basis for the space of 3 × 3 symmetric matrices.
Sol: A 3 × 3 symmetric matrix has the form:
 
a11 a12 a13
A = a12 a22 a23 
a13 a23 a33

where aij = aji for all i, j.

The matrix is completely determined by 6 independent entries: a11 , a12 , a13 , a22 , a23 , a33 .
We can write any 3 × 3 symmetric matrix as:
     
1 0 0 0 1 0 0 0 1
A = a11 0 0 0 + a12 1 0 0 + a13 0 0
     0 (8)
0 0 0 0 0 0 1 0 0
     
0 0 0 0 0 0 0 0 0
+ a22 0 1 0 + a23 0 0 1 + a33 0
     0 0 (9)
0 0 0 0 1 0 0 0 1

Therefore, a basis for the space of 3 × 3 symmetric matrices is:

     

 1 0 0 0 1 0 0 0 1 
E11 = 0 0 0 , E12 = 1 0 0 , E13 = 0 0 0
0 0 0 0 0 0 1 0 0
 

      
 0 0 0 0 0 0 0 0 0 
∪ E22 = 0 1 0 , E23 = 0 0 1 , E33 = 0 0 0
0 0 0 0 1 0 0 0 1
 

Answer: The dimension of the space of 3 × 3 symmetric matrices is 6 , and the

basis consists of the 6 matrices shown above.
Note: In general, the space of n × n symmetric matrices has dimension n(n+1)
2
,
n(n−1)
since we have n diagonal entries and 2 entries above the diagonal that can be
chosen freely.

12. (Treil 1.2.6) Is it possible that vectors v1 , v2 , v3 are linearly dependent, but the
vectors w1 = v1 + v2 , w2 = v2 + v3 , and w3 = v3 + v1 are linearly independent?

8
Sol: We have the transformation:

w1 = v1 + v2 (1)
w2 = v2 + v3 (2)
w3 = v3 + v1 (3)

We can solve this system to express v1 , v2 , v3 in terms of w1 , w2 , w3 :

Adding all three equations: w1 + w2 + w3 = 2(v1 + v2 + v3 )
From the original system:

w1 − w2 + w3 = (v1 + v2 ) − (v2 + v3 ) + (v3 + v1 ) = 2v1 (4)

w1 + w2 − w3 = (v1 + v2 ) + (v2 + v3 ) − (v3 + v1 ) = 2v2 (5)
−w1 + w2 + w3 = −(v1 + v2 ) + (v2 + v3 ) + (v3 + v1 ) = 2v3 (6)

Therefore:
1
v1 = (w1 − w2 + w3 ) (7)
2
1
v2 = (w1 + w2 − w3 ) (8)
2
1
v3 = (−w1 + w2 + w3 ) (9)
2

Since v1 , v2 , v3 are linearly dependent, there exist scalars α1 , α2 , α3 (not all zero)
such that:
α1 v1 + α2 v2 + α3 v3 = 0

Substituting our expressions:

1 1 1
α1 · (w1 − w2 + w3 ) + α2 · (w1 + w2 − w3 ) + α3 · (−w1 + w2 + w3 ) = 0 (10)
2 2 2

Simplifying:
1
[(α1 + α2 − α3 )w1 + (−α1 + α2 + α3 )w2 + (α1 − α2 + α3 )w3 ] = 0
2

This gives us:

(α1 + α2 − α3 )w1 + (−α1 + α2 + α3 )w2 + (α1 − α2 + α3 )w3 = 0

 
1 1 −1
Since the transformation matrix A = −1 1 1  has det(A) = −4 ̸= 0, it
1 −1 1
maps any non-zero vector (α1 , α2 , α3 ) to a non-zero vector.
Therefore, the coefficients of w1 , w2 , w3 cannot all be zero, proving that w1 , w2 , w3
are linearly dependent.
Answer: No, it is not possible

9
13. (Treil 1.2.5) Let a system of vectors v1 , v2 , . . . , vr be linearly independent but not
generating (i.e., not spanning the entire space). Show that it is possible to find a
vector vr+1 such that the system v1 , v2 , . . . , vr , vr+1 is linearly independent.
Hint:
Pr Take for vr+1 any vector that cannot be represented as a linear combination
k=1 k vk and show that the system v1 , v2 , . . . , vr , vr+1 is linearly independent.
α
Sol:

Theorem 1. If {v1 , v2 , . . . , vr } is a linearly independent set that does not span the
vector space V , then there exists a vector vr+1 ∈ V such that {v1 , v2 , . . . , vr , vr+1 }
is linearly independent.

Proof. Step 1: Understanding the Given Conditions

Given:

ˆ The vectors v1 , v2 , . . . , vr are linearly independent

ˆ The system is not generating, meaning span{v1 , v2 , . . . , vr } ̸= V

Step 2: Existence of vr+1

Since {v1 , v2 , . . . , vr } does not span V , there exists at least one vector in V that
cannot be expressed as a linear combination of v1 , v2 , . . . , vr .
Let vr+1 be any such vector, i.e., vr+1 ∈
/ span{v1 , v2 , . . . , vr }.
This means that the equation:
r
X
vr+1 = αk vk
k=1

has no solution for any scalars α1 , α2 , . . . , αr .

Step 3: Proving Linear Independence
We need to prove that {v1 , v2 , . . . , vr , vr+1 } is linearly independent.
Consider the linear combination:

β1 v1 + β2 v2 + · · · + βr vr + βr+1 vr+1 = 0

We need to show that this implies β1 = β2 = · · · = βr = βr+1 = 0.

Step 4: Case Analysis
Case 1: βr+1 = 0
If βr+1 = 0, then our equation becomes:

β1 v1 + β2 v2 + · · · + βr vr = 0

Since {v1 , v2 , . . . , vr } are linearly independent by assumption, this implies:

β1 = β2 = · · · = βr = 0

Case 2: βr+1 ̸= 0

10
 If βr+1 ̸= 0, we can rearrange the equation:
βr+1 vr+1 = −(β1 v1 + β2 v2 + · · · + βr vr )

Dividing by βr+1 :
β1 β2 βr
vr+1 = − v1 − v2 − · · · − vr
βr+1 βr+1 βr+1

This would mean that vr+1 can be expressed as a linear combination of v1 , v2 , . . . , vr ,

i.e., vr+1 ∈ span{v1 , v2 , . . . , vr }.
But this contradicts our choice of vr+1 which was specifically chosen to NOT be in
the span of {v1 , v2 , . . . , vr }.
Step 5: Conclusion
Since Case 2 leads to a contradiction, we must have βr+1 = 0.
From Case 1, when βr+1 = 0, we get β1 = β2 = · · · = βr = 0.
Therefore:
β1 = β2 = · · · = βr = βr+1 = 0

This proves that the system {v1 , v2 , . . . , vr , vr+1 } is linearly independent.

This result is fundamental in linear algebra and forms the basis for extending any
linearly independent set to a basis of the vector space. The process can be continued
until we obtain a spanning set, thus constructing a basis.
14. (Treil 1.3.3.(d)) For the linear transformation below find the matrix of T .
T : Pn → Pn , T (f (t)) = 2f (t) + 3f ′ (t) − 4f ′′ (t)
with respect to the standard basis {1, t, t2 , . . . , tn }.
Sol: Let T : Pn → Pn be defined by:
T (f (t)) = 2f (t) + 3f ′ (t) − 4f ′′ (t).
We compute the matrix representation of T with respect to the standard basis
B = {1, t, t2 , . . . , tn }.

Step 1: Apply T to the basis vectors

For f (t) = tk , we have:
f ′ (t) = ktk−1 , f ′′ (t) = k(k − 1)tk−2 .
Thus,
T (tk ) = 2tk + 3ktk−1 − 4k(k − 1)tk−2 .
This holds for k ≥ 2. For k = 0 and k = 1:
T (1) = 2 · 1 + 3 · 0 − 4 · 0 = 2,
T (t) = 2t + 3 · 1 − 4 · 0 = 3 + 2t.

11
Step 2: Express T (tk ) in terms of the basis
For each k, write T (tk ) as a linear combination of 1, t, t2 , . . . , tn :
ˆ For k = 0: T (1) = 2 · 1.
ˆ For k = 1: T (t) = 3 · 1 + 2 · t.
ˆ For k = 2: T (t2 ) = −8 · 1 + 6 · t + 2 · t2 .
ˆ For k ≥ 2:
T (tk ) = −4k(k − 1)tk−2 + 3ktk−1 + 2tk .

Step 3: Construct the matrix

The matrix A of T is an (n + 1) × (n + 1) matrix. The columns are the coordinate
vectors of T (tk ) for k = 0, 1, . . . , n.
Let the rows and columns be indexed from 0 to n. Then:
ˆ Column 0 (corresponding to T (1)):
[2, 0, 0, . . . , 0]T .
ˆ Column 1 (corresponding to T (t)):
[3, 2, 0, . . . , 0]T .
ˆ For j ≥ 2, column j (corresponding to T (tj )) has:
aj−2,j = −4j(j − 1),
aj−1,j = 3j,
aj,j = 2,
and all other entries are zero.

Step 4: Example for n = 2

For n = 2, the basis is {1, t, t2 }. The matrix is:
 
2 3 −8
A = 0 2 6  .
0 0 2

Step 5: General matrix

The general (n + 1) × (n + 1) matrix A is upper triangular and given by:
 
2 3 −4 · 2 · 1 −4 · 3 · 2 · · · −4n(n − 1)
0 2
 3·2 −4 · 3 · 2 · · · 3n 

0 0 2 3 · 3 · · · −4n(n − 1) 
A = 0 0 .
 
 0 2 · · · 3n 
 .. .. .. .. .. .. 
. . . . . . 
0 0 0 0 ··· 2

12
More precisely, for j = 0, 1, . . . , n:

a0,0 = 2,
a0,1 = 3, a1,1 = 2,
For j ≥ 2 : aj−2,j = −4j(j − 1), aj−1,j = 3j, aj,j = 2,

and all other entries are zero.

13
Change of Basis Matrices
Rule of thumb: from S to B express vectors in S with respect to vectors in B.

1. Find the change of basis matrix from S to B, i.e., [I]BS , given two bases:
S = {(1, 0), (0, 1)}T and B = {(1, 2), (2, 1)}T .
Sol. From S to B:
For S → B, we will make use of one result derived earlier, i.e.
 
−1 a11 a12
[I]BS = B = T =
←− a21 a22

Using option 1: We express vectors in S with respect to vectors in B. So,

Step 1:         
1 1 2 1 2 a11
= a11 + a21 =
0 2 1 2 1 a21
ˆ Look what does it mean? :
Column 1 of T multiplied by [B] and produces col. 1 of [S].
   
a11 −1 1
=B
a21 0
Step 2:         
0 1 2 1 2 a12
= a12 + a22 =
1 2 1 2 1 a22
   
a12 −1 0
=B
a22 1
   −1
a11 a12 1 2
Combining the two cases above: [I]BS =T = = .I = B −1
a21 a22 2 1
From B to S
Now, for B → S: express columns of B with respect to columns of S. So,
       
1 1 0 b
= b11 + b21 = I2 11
2 0 1 b21
     
b11 −1 1 1
⇒ =I =
b21 2 2
Similarly,      
b12 −1 2 2
=I =
b22 1 1
Therefore,    
b11 b12 1 2
[I]SB = = =B
b21 b22 2 1

◦◦ Now, say, another basis A = {(1, 0), (1, 1)}T .

14
 Can we go from B → A? i.e., we need [I]AB . We can use the ”via” formula.
←−

[I]AB = [I]AS · [I]SB

←− ←− ←−

=⇒ [I]AB = A−1 · B
A Verification of
 the
 formula:    
3 3 1
Suppose ∃ v = ∈ R2 . Thus, we have [v]S = but [v]B = .
3 3 1
Find [I]BS so that [I]BS [v]S = [v]B .
2. (Treil. ) In a space of polynomials of degree at most 1, bases are supposed to be
A = {1, 1 + x} and B = {1 + 2x, 1 − 2x}.
In a matrix
 form, these
 bases
 can be written as follows:
1 1 1 1
A= ,B= . Note: P1 = R2 .
0 1 2 −2
We want to find the change of coordinate matrix [I]BA .
Option 1: Simplest method via standard basis
Find [I]BA via the standard basis S = {1, x} in P1 , which corresponds to
S = {(1, 0), (0, 1)}T in R2 .

(a) Find [I]SA and [I]SB . By definition, these matrices are simply the basis ma-
trices themselves.
   
1 1 1 1
[I]SA = A = and [I]SB = B =
0 1 2 −2
(b) Find the inverse matrices needed for the transformation.
 
−1 1 −1
[I]AS = A =
0 1
     
−1 1 −2 −1 1 −2 −1 1/2 1/4
[I]BS = B = =− =
(1)(−2) − (1)(2) −2 1 4 −2 1 1/2 −1/4
(c) Use the formula for change of basis through an intermediate basis.
[I]BA = [I]BS · [I]SA = B −1 A
    
1/2 1/4 1 1 (1/2)(1) + (1/4)(0) (1/2)(1) + (1/4)(1)
[I]BA = =
1/2 −1/4 0 1 (1/2)(1) + (−1/4)(0) (1/2)(1) + (−1/4)(1)
 
1/2 3/4
[I]BA =
1/2 1/4

Option 2: Direct decomposition of vectors

For [I]BA , take the vectors from basis A and decompose them with respect to basis
B. The resulting coordinate vectors
  will form the
 columns of the matrix [I]BA .
1 1
Let the columns of A be A1 = and A2 = .
0  1 
1 1
Let the columns of B be B1 = and B2 = .
2 −2

15
 (a) Find the coordinates of A1 with respect to B.
    
1 1 1 a11
A1 = a11 B1 + a21 B2 =⇒ =
0 2 −2 a21
Solving for the coefficients:
        
a11 −1 1 1/2 1/4 1 1/2
=B = =
a21 0 1/2 −1/4 0 1/2

(b) Find the coordinates of A2 with respect to B.

    
1 1 1 a12
A2 = a12 B1 + a22 B2 =⇒ =
1 2 −2 a22
Solving for the coefficients:
          
a12 −1 1 1/2 1/4 1 1/2 + 1/4 3/4
=B = = =
a22 1 1/2 −1/4 1 1/2 − 1/4 1/4

(c) Combine the column vectors to form the matrix [I]BA .

 
1/2 3/4
[I]BA =
1/2 1/4

3. Exercise: Consider a linear transformation, T : V → V , so that T (v) = −2v.

Find the matrix for T in the standard basis S and then in basis B, where B =
{(2, 3), (1, 2)}T .

Sol. The standard basis is S = {(1, 0), (0, 1)}T and the new basis is B = {(2, 3), (1, 2)}T .
◦◦ The relationship between a vector’s coordinates in different bases is given by the
change of basis formula: [v]B = [T (v)]B = [I]BS [v]S .
◦◦ The relationship for a transformation matrix in different bases is given by:

[T ]BB = [I]BS · [T ]SS · [I]SB

←− ←−

Step 1: Find the Matrix of TSS in the Standard Basis S

We apply the transformation T to each basis vector in S and express the result as
a linear combination of the vectors in S.

T (1, 0)T = −2 · (1, 0)T = −2 · (1, 0)T + 0 · (0, 1)T

T (0, 1)T = −2 · (0, 1)T = 0 · (1, 0)T − 2 · (0, 1)T

The columns of the matrix for T in basis S, denoted [T ]SS , are the coordinate
vectors of the transformed basis vectors.
 
−2 0
[T ]SS =
0 −2

Step 2: Find the Matrix of TBB in Basis B

We apply the transformation T to each basis vector in B and express the result as
a LC of the vectors in B. Let the basis vectors be v1 = (2, 3)T and v2 = (1, 2)T .

16
 For the first column:

T (v1 ) = T (2, 3)T = −2 · (2, 3)T = (−4, −6)T

We need to find coefficients a11 and a21 such that:

        
−4 2 1 2 1 a11
= a11 + a21 =
−6 3 2 3 2 a21
 
2 1
The change of basis matrix from B to S is [I]SB = . To solve for the
3 2
coefficients, we need its inverse:
 −1  
−1 2 1 2 −1
[I]BS = ([I]SB ) = =
3 2 −3 2

Multiplying by the inverse:

          
a11 2 −1 −4 (2)(−4) + (−1)(−6) −8 + 6 −2
= = = =
a21 −3 2 −6 (−3)(−4) + (2)(−6) 12 − 12 0

For the second column:

T (v2 ) = T (1, 2) = −2(1, 2) = (−2, −4)

We need to find coefficients a12 and a22 such that:

     
−2 2 1
= a12 + a22
−4 3 2

Multiplying by the inverse of [I]SB :

          
a12 2 −1 −2 (2)(−2) + (−1)(−4) −4 + 4 0
= = = =
a22 −3 2 −4 (−3)(−2) + (2)(−4) 6−8 −2

Step 3: Form the Matrix [T ]BB The matrix for the linear transformation T in
basis B, denoted [T ]B , is composed of the coordinate vectors found in the previous
step.    
a11 a12 −2 0
[T ]BB = =
a21 a22 0 −2

Conclusion The matrix for the linear transformation T is the same in both the
standard basis S and basis B. This is expected for a scalar transformation, as it
scales all vectors by the same factor, regardless of the coordinate system.
2
4. Treil 2.3.4: Let T be
 a linear
 transformation (LT) in the standard basis in R
x 3x + y
defined by T = , where x, y ∈ R. Find the matrix of T first in the
y x − 2y
standard basis (S) and then in basis B = {(1, 1), (1, 2)}T .

17
Sol. Step 1: Matrix of T in the standard basis S
The linear transformation T can be represented by a matrix. We can find this
matrix by inspection also from the given definition.
      
x 3x + y 3 1 x
T = =
y x − 2y 1 −2 y
 
3 1
Thus, the matrix for T in the standard basis is [T ]SS = := A.
1 −2
Step 2: Find the matrix of transformation in basis B
Recall that the relationship between the matrix of a transformation in two different
bases is given by:
[T ]BB = [I]BS · [T ]SS · [I]SB
where [I]SB is the change of basis matrix from B to S.
We can also find [T ]BB by a direct method.
This method involves expressing the columns of the standard basis matrix A = [T ]SS
in terms of the new basis vectors in B.
Let’s express the basis vectors of S in terms of the basis vectors of B.
The basis B is given by {v1 , v2 }T = {(1, 1), (1, 2)}T . Let’s find the change of basis
matrix [I]SB . The columns of this matrix are the coordinates of the vectors of B
expressed in the standard basis.
 
1 1
[I]SB ≡ B =
1 2

We also need the inverse, [I]BS :

   
−1 1 2 −1 2 −1
[I]BS = B = =
(1)(2) − (1)(1) −1 1 −1 1

Now, using the change of basis formula:

   
2 −1 3 1 1 1
[T ]BB = [I]BS · [T ]SS · [I]SB =
−1 1 1 −2 1 2

Let’s calculate the product. First, [I]BS [T ]SS :

      
2 −1 3 1 (2)(3) + (−1)(1) (2)(1) + (−1)(−2) 5 4
= =
−1 1 1 −2 (−1)(3) + (1)(1) (−1)(1) + (1)(−2) −2 −3

Next, multiply the result by [I]SB :

      
5 4 1 1 (5)(1) + (4)(1) (5)(1) + (4)(2) 9 13
= =
−2 −3 1 2 (−2)(1) + (−3)(1) (−2)(1) + (−3)(2) −5 −8
So, the matrix of T in basis B is:
 
9 13
[T ]BB =
−5 −8

18
5. 
(Exercise:)
 Consider a linear transformation T : R2 → R2 with matrix [T ]BA =
1 2
, where A and B are two bases of R2 .
3 4
Let A′ = {(1, 2), (3, 4)} and B ′ = {(1, 0), (0, 1)}. Find the matrix representation of
T in the new bases, denoted as [T ]B′ A′ .
Solution: The relationship between the matrix of a linear transformation in
different bases is given by the formula:
[T ]B′ A′ = [I]B′ B · [T ]BA · [I]AA′

Here, [I]AA′ is the change of basis matrix from A′ to A, and [I]B′ B is the change of
basis matrix from B to B ′ .

Step 1: Find the change of basis matrices

We need to find the matrices [I]AA′ and [I]B′ B . We are given that A and B are the
standard bases for their respective spaces.
ˆ The matrix [I]AA′ is found by expressing the basis vectors of A′ in terms of the
basis vectors of A. Since A is the standard basis, the columns of this matrix
are simply the vectors of A′ .
 
1 3
[I]AA′ =
2 4
ˆ The matrix [I]B′ B is found by expressing the basis vectors of B in terms
of the basis vectors of B ′ . Let’s assume the previous example’s basis B =
{(1, 1), (1, −1)}. Since B ′ is the standard basis, the columns of this matrix are
the vectors of B.  
1 1
[I]B′ B =
1 −1

Step 2: Calculate the matrix [T ]B′ A′

Now, we multiply the matrices in the correct order as per the formula:
   
1 1 1 2 1 3
[T ]B′ A′ =
1 −1 3 4 2 4
First, let’s calculate the product of the last two matrices:
      
1 2 1 3 (1)(1) + (2)(2) (1)(3) + (2)(4) 5 11
= =
3 4 2 4 (3)(1) + (4)(2) (3)(3) + (4)(4) 11 25
Next, multiply this result by the first matrix:
      
1 1 5 11 (1)(5) + (1)(11) (1)(11) + (1)(25) 16 36
[T ]B′ A′ = = =
1 −1 11 25 (1)(5) + (−1)(11) (1)(11) + (−1)(25) −6 −14
The matrix representation of the linear transformation T from basis A′ to basis B ′
is:  
16 36
[T ]B′ A′ =
−6 −14

19