Linear Algebra

Notes taken by Jack Chen

Jul 9, 2022
Contents

I 2
1 Linear Transformations and Matrices 3
1.1 dual spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Elementary Matrix Operations and Systems of Linear Equations 12

2.1 Elementary Matrix Operations and Elementary Matrices . . . . . . . . . . . . . . 12
2.2 The Rank of a Matrix and Matrix Inverses . . . . . . . . . . . . . . . . . . . . . 13
2.3 Systems of Linear Equations - Theoretical Aspects . . . . . . . . . . . . . . . . . 14
2.4 Systems of Linear Equations - Computational Aspects . . . . . . . . . . . . . . . 15

3 Determinants 17
3.1 Determinant of Order 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2 Determinant of Order n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.3 Properties of Determinants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Index 33

1
Part I

2
Chapter 1

Linear Transformations and

Matrices

1.1 dual spaces

In this section, linear transformations from a vector space V into its field of scalars F , which is
itself a vector space of dimension 1 over F , are focused. Such a linear transformation is called a
linear functional on V .

Example. Let V be the vector space of continuous real-valued functions on the interval
[0, 2π]. Fix a function g ∈ V . The function h : V → R defined by
Z 2π
1
h(x) = x(t)g(t) dt
2π 0

is a linear functional on V .

Definition 1. For a vector space V over F , the dual space of V is defined to be L(V, F ),
denoted by V ∗

Note. If V is a finite-dimensional vector space over F , then

dim(V ∗ ) = dim(L(V, F )) = dim(V ) · dim(F ) = dim(V ).

Hence V and V ∗ are isomorphic.

Since V and V ∗ are isomorphic (if they are finite-dimensional), the next is the relation between
bases for V and V ∗ .

Theorem 1. Suppose that V is a finite-dimensional vector space with the ordered basis
β = {x1 , x2 , . . . , xn }. Let fi (1 ≤ i ≤ n) be the ith coordinate function with respect to β,
and let β ∗ = {f1 , f2 , · · · , fn }. Then β ∗ is an ordered basis for V ∗ , and, for any f ∈ V ∗ , we

3
Linear Algebra 1.1. DUAL SPACES

have
n
X
f= f (xi )fi .
i=1

Proof. Let f ∈ V ∗ . We need to only show that

n
X
f= f (xi )fi
i=1

from which it follows that β ∗ generates V ∗ , since dim(V ∗ ) = n. For 1 ≤ j ≤ n, we have

n
! n
X X
f (xi )fi (xj ) = f (xi )fi (xj )
i=1 i=1
n
X
= f (xi )δij = f (xj ).
i=1
Pn
Therefore f = i=1 f (xi )fi .

Definition 2. Suppose that V is a finite-dimensional vector space with the ordered basis
β = {x1 , x2 , . . . , xn }. The ordered basis β ∗ = {f1 , f2 , . . . , fn } of V ∗ that satisfies fi (xj ) = δij
(1 ≤ i, j ≤ n) the dual basis of β.

Example. Let β = {(2, 1), (3, 1)} be an ordered basis for R2 . Suppose that the dual basis
of β is given by β ∗ = {f1 , f2 }. To explicitly determine a formula for f1 , we consider the
equations

1 = f1 (2, 1) = f1 (2e1 + 1e2 ) = 2f1 (e1 ) + f1 (e2 )

0 = f1 (3, 1) = f1 (3e1 + 1e2 ) = 3f1 (e1 ) + f1 (e2 ),

and

0 = f2 (2, 1) = f2 (2e1 + 1e2 ) = 2f2 (e1 ) + f2 (e2 )

1 = f2 (3, 1) = f2 (3e1 + 1e2 ) = 3f2 (e1 ) + f2 (e2 ).

Solving the equations, we obtain f1 (e1 ) = −1, f1 (e2 ) = 3, f2 (e1 ) = 1, and f2 (e2 ) = −2; that
is, f1 (x, y) = −x + 3y and f2 (x, y) = x − 2y, i.e.,

f1 = −g1 + 3g2
f2 = g1 + 2g2 ,

where {g1 , g2 } is the dual basis of the standard ordered basis for R2 .

Theorem 2. Let V and W be finite-dimensional vector spaces over F with the ordered basis β
and γ, respectively. For any linear transformation T : V → W , the mapping T t : W ∗ → V ∗

CHAPTER 1. LINEAR TRANSFORMATIONS AND MATRICES 4

Linear Algebra 1.1. DUAL SPACES

defined by T t (g) = gT for all g ∈ W ∗ is a linear transformation with the property that
∗
[T t ]βγ ∗ = ([T ]γβ )t .

Proof. For g ∈ W ∗ , T t (g) = gT is a linear functional on V and hence is in V ∗ . Thus T t

maps W ∗ to V ∗ . Given g1 , g2 ∈ W ∗ . Then

T t (g1 + rg2 ) = (g1 + rg2 )T

= g1 T + (rg2 )T
= T t (g1 ) + T t (rg2 );

it follows that T t is linear.

Let β = {x1 , x2 , . . . , xn } and γ = {y1 , y2 , . . . , ym } with the dual bases β ∗ = {f1 , f2 , . . . , fn }
∗
and γ ∗ = {g1 , g2 , . . . , gm }, respectively. To find jth column of [T t ]βγ ∗ , we express T t (gj ) as a
linear combination of the vectors of β ∗ . Note that
n
X
T t (gj ) = gj T = (gj T )(xi )fi .
i=1

∗
So the row i, column j entry of [T t ]βγ ∗ is
m
X
(gj T )(xi ) = gj (T (xi )) = gj ([T ]γβ )ki yk
k=1
m
X
= ([T ]γβ )ki gj yk
k=1
Xm
= ([T ]γβ )ki δjk
k=1
= ([T ]γβ )ji .

Hence ∗
[T t ]βγ ∗ = ([T ]γβ )t .

The linear transformation T t defined in the previous theorem is called the transpose of T .

Example. Define T : P1 (R) → R2 by T (p(x)) = (p(0), p(2)). Let β and γ be the standard
ordered bases for P1 (R) and R2 , respectively. Clearly,
 
1 0
[T ]γβ = .
1 2
∗
We compute [T t ]βγ ∗ directly form the definition. Let β ∗ = {f1 , f2 } and γ ∗ = {g1 , g2 }.

CHAPTER 1. LINEAR TRANSFORMATIONS AND MATRICES 5

Linear Algebra 1.1. DUAL SPACES

 
∗ a b
Suppose that [T t ]βγ ∗ = . Then
c d

T t (g1 )(1) = (af1 + cf2 )(1) = af1 (1) + cf2 (1) = a(1) + c(0) = a,
T t (g1 )(x) = (af1 + cf2 )(x) = af1 (x) + cf2 (x) = a(0) + c(1) = c.

But also

T t (g1 )(1) = g1 (T (1)) = g1 (1, 1) = 1,

T t (g1 )(x) = g1 (T (x)) = g1 (0, 2) = 0.

So a = 1 and c = 0. Similarly,

T t (g2 )(1) = (bf1 + df2 )(1) = bf1 (1) + df2 (1) = b(1) + d(0) = b,
T t (g2 )(x) = (bf1 + df2 )(x) = bf1 (x) + df2 (x) = b(0) + d(1) = d;

and

T t (g2 )(1) = g2 (T (1)) = g2 (1, 1) = 1,

T t (g2 )(x) = g1 (T (x)) = g2 (0, 2) = 2.

So b = 1 and d = 2. Hence it yields

 
∗ 1 1
[T t ]γβ∗ = = ([T ]γβ )t ,
0 2

as predicted by the previous theorem.

We now concern that any finite-dimensional vector space V can be identified in a natural way
with its double dual V ∗∗ .
For a vector x ∈ V , we define x̂ : V ∗ → F by x̂(f ) = f (x) for all f ∈ V ∗ . It is easy to verify
that x̂ is a linear functional on V ∗ , so x̂ ∈ V ∗∗ . The correspondence x ↔ x̂ allows us to define
the desired isomorphism between V and v ∗∗ .

Lemma. let V be a finite-dimensional vector space, and let x ∈ V . If x̂(f ) = 0 for all f ∈ V ∗ ,
then x = 0.

Proof. Let x 6= 0. Choose an ordered basis β = {x1 , x2 , . . . , xn } for V such that x1 = x. Let
{f1 , f2 , . . . , fn } be the dual basis of β. Then x̂(f1 ) = f1 (x) = 1 6= 0.

Theorem 3. Let V be a finite-dimensional vector space, and define ψ : V → V ∗∗ by ψ(x) = x̂.

Then ψ is an isomorphism.

Proof. 1° ψ is linear: Let x, y ∈ V and c ∈ F . For f ∈ V ∗ , we have

ψ(cx + y)(f ) = f (cx + y) = cf (x) + f (y) = cx̂(f ) + ŷ(f )

= (cx̂ + ŷ)(f ) = (cψ(x) + ψ(y))(f ).

CHAPTER 1. LINEAR TRANSFORMATIONS AND MATRICES 6

Linear Algebra 1.1. DUAL SPACES

Therefore
ψ(cx + y) = (cψ(x) + ψ(y)).

2° ψ is one-to-one: Let x ∈ V s.t. ψ(x) = 0. Then 0 = ψ(x)(f ) = f (x) for all f ∈ V ∗ . By

the previous lemma, we conclude that x = 0.
3° ψ is an isomorphism: This follows from 2° and the fact that dim(V ) = dim(V ∗∗ ).

Corollary. Let V be a finite-dimensional vector space with dual space V ∗ . Then every ordered
basis for V ∗ is the dual basis for some basis for V .

Proof. Let {f1 , f2 , . . . , fn } be an ordered basis for V ∗ . By the previous theorems, for
this basis there exists a dual basis, say {g1 , g2 , . . . , gn }, in V ∗∗ which satisfies gi (fj ) =
δij x̂i (fi ) = fj (xj ) for all i and j. Thus gi = x̂i for all i and {f1 , f2 , . . . , fn } is the dual basis
of {x1 , x2 , . . . , xn }.

Exercise. Prove that a function T : F n → F m is linear if and only if there exist f1 , f2 , . . . , fm ∈

(F n )∗ such that T (x) = (f1 (x), f2 (x), . . . , fm (x)) for all x ∈ F n .

Proof. Suppose that T is linear. Define fi = gi T ∈ (F n )∗ for x ∈ F n ; that is, fi = T t (gi )

for 1 ≤ i ≤ m, where {g1 , g2 , . . . , gm } is the dual basis of the standard ordered basis for F m .
Then fi is linear for each i, and

T (x) = (g1 T (x), g2 T (x), . . . , gm T (x)).

Conversely, let {ei }i=1,2,...,m be the standard basis for F m . We have

m
X
T (x) = fi (x)ei
i=1

which follows that T is linear since each fi is a linear functional.

Exercise. Let V = Pn (F ), and let c0 , c1 , . . . , cn be distinct scalars in F .

1. For 0 ≤ i ≤ n, define fi ∈ V ∗ by fi (p(x)) = p(ci ). Prove that {f0 , f1 , . . . , fn } is a basis
for V ∗ .
2. Show that there exist unique polynomials p0 (x), p1 (x), . . . , pn (x) such that pi (ci ) = δij
for 0 ≤ i ≤ n. These polynomials are the Lagrange polynomials.
3. For any scalars a0 , a1 , . . . , an (not necessarily distinct), deduce that there exists a
unique polynomial q(x) of degree at most n such that q(ci ) = ai for 0 ≤ i ≤ n. In fact,
n
X
q(x) = ai pi (x).
i=0

CHAPTER 1. LINEAR TRANSFORMATIONS AND MATRICES 7

Linear Algebra 1.1. DUAL SPACES

4. Deduce the Lagrange interpolation formula:

n
X
p(x) = p(ci )pi (x)
i=0

Proof. 1. For each i, let p1 , p2 ∈ Pn (F ) and r ∈ F , we have

fi ((p1 + rp2 )(x)) = (p1 + rp2 )(ci )

= p1 (ci ) + rp2 (ci )
= fi (p1 (x)) + rfi (p2 (x)).

Thus fi ∈ V ∗ for all i. Since #{f0 , f2 , . . . , fn } = n + 1 = dim(V ), it remains to show

that {fi }ni=0 is linearly independent. Let r0 , r1 , . . . , rn ∈ F such that
n
X
ri fi = O
i=0
Q
the zero linear functional. Choose pj (x) = k6=j (x − ck ) for 0 ≤ j ≤ n. Then, for
0 ≤ i, j ≤ n, fi (pj (x)) = pj (ci ) = 0 if i 6= j and fi (pj (x)) 6= 0 if i = j since ci are
distinct. Moreover,

0 = O(pj (x))
n
!
X
= ri fi pj (x)
i=0
n
X
= ri fi (pj (x))
i=0
Xn
= ri pj (ci )
i=0
= rj pj (cj );

it follows that rj = 0. Thus {fi } is linearly independent and hence a basis for V ∗ .

2. By the previous theorem, there exists a basis

β = {p0 (x), p1 (x), . . . , pn (x)}

for V such that {f0 , f1 , . . . , fn } is its dual basis and that pj (ci ) = fi (pj ) = δij .
For each j, let qj ∈ Pn (F ) s.t. qj (ci ) = δij , we have
n
X
δij = qj (ci ) = rk pk (ci ) = ri
k=0

where all rk ∈ F . So qj = pj for all j and are unique.

CHAPTER 1. LINEAR TRANSFORMATIONS AND MATRICES 8

Linear Algebra 1.1. DUAL SPACES

3. Since {p0 , p1 , . . . , pn } is a basis for V ∗ ,

n
X
q(x) = ri pi (x)
i=0

where all ri ∈ F . Then ai = q(ci ) = ri for all i.

4. From the result of the above,
n
X n
X
q(x) = ai pi (x) = q(ci )pi (x),
i=0 i=0

where pi (x) ∈ Pn (F ) satisfies pi (cj ) = δij .

Exercise. Let V and W be finite-dimensional vector spaces over F , and let ψ1 and ψ2 be the
isomorphisms between V and V ∗∗ and W and W ∗∗ , respectively, as defined in the preceding
theorem. Let T : V → W be linear, and define T tt = (T t )t . Prove that ψ2 T = T tt ψ1 .

Proof. Let v ∈ V . We have

ˆ ∈ W ∗∗
(ψ2 T )(v) = ψ2 (T (v)) = (T (v))

and
(T tt ψ1 )(v) = T tt (ψ1 (v)) = T tt v̂ = v̂T t ∈ W ∗∗ .
ˆ = T tt v̂, let f ∈ W ∗ . Then
To verify that (T (v))

ˆ
(T (v))(f ) = f (T (v))

and
(v̂T t )(f ) = v̂(T t (f )) = v̂(f T ) = (f T )(v) = f (T (v));
ˆ = T tt v̂. This completes the proof.
hence (T (v))

Exercise. Let V be a finite-dimensional vector space with the ordered basis β. Prove that
ψ(β) = β ∗∗ , where ψ is defined in the preceding theorem.

Proof. Let β = {x1 , x2 , . . . , xn }, and let β ∗ = {f1 , f2 , . . . , fn } and β ∗∗ = {F1 , F2 , . . . , Fn } be the

dual bases of β and β ∗ , respectively. Then, for 1 ≤ i, j ≤ n,

ψ(xi )(fj ) = x̂i (fj ) = fj (xi ) = δij

and
Fi (fj ) = δij ;
thus Fi = ψ(xi ). Hence we obtain ψ(β) = β ∗∗ .

CHAPTER 1. LINEAR TRANSFORMATIONS AND MATRICES 9

Linear Algebra 1.1. DUAL SPACES

Exercise. Let V be a finite-dimensional vector space over F . For every subset S of V , define
the annihilator S 0 of S as

S 0 = {f ∈ V ∗ | f (x) = 0 for all x ∈ S}

1. Prove that S 0 is a subspace of V ∗ .

/ W , prove that there exists f ∈ W 0 such that f (x) 6= 0.

2. If W is a subspace of V and x ∈
3. Prove that (S 0 )0 = span(ψ(S)), where ψ is defined as in the preceding theorem.
4. For subspaces W1 and W2 , prove that W1 = W2 if and only if W10 = W20 .

5. For subspaces W1 and W2 , show that (W1 + W2 )0 = W10 ∩ W20 .

Proof. 1. Clearly, the zero function is in S 0 . Let f, g ∈ S 0 and r ∈ F . Then, for x ∈ S,

(f + rg)(x) = f (x) + rg(x) = 0 + r · 0 = 0;

thus f + rg ∈ S 0 . Hence S 0 ≤ V ∗ .
2. Let β = {x1 , x2 , . . . , xk } be an ordered basis for W . Since x ∈
/ W , we can extend β to
an ordered basis β 0 = {x1 , x2 , . . . , xn } (n > k) for V which satisfies x = xi for some
i > k. Fix i, let f be the ith coordinate function of β 0 .
3. (a) S 0 = (span(S))0 : Let f ∈ S 0 . For x ∈ span(S),

x = r1 x1 + r2 x2 + · · · + rk xk

for some xi ∈ S and ri ∈ F . Then

k
! k k
X X X
f (x) = f ri xi = ri f (xi ) = (ri · 0) = 0.
i=1 i=1 i=1

So f ∈ (span(S))0 and hence S 0 ⊆ (span(S))0 .

Let f ∈ (span(S))0 . Then f (x) = 0 for all x ∈ S since S ⊆ span(S). So f ∈ S 0 ,
(span(S))0 ⊆ S 0 , and hence (span(S))0 = S 0 .
(b) span(ψ(S)) = ψ(span(S)): It is clearly since ψ is an isomorphism.
(c) ((span(S))0 )0 = ψ(span(S)): Let x̂ ∈ ((span(S))0 )0 . If x ∈
/ span(S), then, form
the above, there exists f ∈ (span(S))0 such that x̂(f ) = f (x) 6= 0; it follows
/ ((span(S))0 )0 (→←). So x ∈ span(S), x̂ = ψ(x) ∈ ψ(span(S)). Hence
that x̂ ∈
((span(S))0 )0 ⊆ ψ(span(S)).
Let x̂ ∈ ψ(span(S)). Then x ∈ span S, and, for f ∈ (span(S))0 ,

x̂(f ) = f (x) = 0.

Thus x̂ ∈ ((span(S))0 )0 and ψ(span(S)) ⊆ ((span(S))0 )0 . Hence ((span(S))0 )0 =

ψ(span(S)).

CHAPTER 1. LINEAR TRANSFORMATIONS AND MATRICES 10

Linear Algebra 1.1. DUAL SPACES

4. If W1 = W2 , then W10 = W20 clearly. Suppose that W10 = W20 . Then

ψ(W1 ) = (W10 )0 = (W20 )0 = ψ(W2 ).

So
W1 = ψ −1 (ψ(W1 )) = ψ −1 (ψ(W2 )) = W2 .

5. Let f ∈ (W1 + W2 )0 . Then for w1 ∈ W1 and w2 ∈ W2 , f (w1 ) = f (w2 ) = 0 since

W1 ⊆ W1 + W2 and W2 ⊆ W1 + W2 . Thus f ∈ W10 ∩ W20 and (W1 + W2 )0 ⊆ W10 ∩ W20 .
Let f ∈ W10 ∩ W20 . Then, for w1 ∈ W2 and w2 ∈ W2 , f (w1 + w2 ) = f (w1 ) + f (w2 ) =
0 + 0 = 0; it follows that f ∈ (W1 + W2 )0 . Thus W10 ∩ W20 ⊆ (W1 + W2 )0 and hence
W10 ∩ W20 = (W1 + W2 )0 .

CHAPTER 1. LINEAR TRANSFORMATIONS AND MATRICES 11

Chapter 2

Elementary Matrix Operations

and Systems of Linear Equations

2.1 Elementary Matrix Operations and Elementary Matri-

ces
Definition 3. Let A be an m × n matrix. Any one of the following three operations on the
row [column] of A is call elementary row [column] operation:
(1) interchanging any two rows [columns] of A;

(2) multiplying any tow row [columns] of A;

(3) adding any scalar multiple of a row [column] of A to another row [column].
Any of these three operations is called an elementary operation. Elementary operations
are of type 1, type 2, or type 3 depending on whether they are obtained by (1), (2), or
(3).

Definition 4. An n×n elementary matrix is a matrix obtained by performing an elementary

operation on In . The elementary matrix is said to be of type 1, 2, or 3 according to
whether the elementary operation performed on In is a type 1, 2, or 3 operation, respectively.

Theorem 4. Let A ∈ Mm×n (F ), and suppose that B is obtained from A by performing

an elementary row [column] operation. Then there exists an m × m [n × n] elementary
matrix E such that B = EA [B = AE]. In fact, E is obtained from Im [In ] by performing
the same elementary row [column] operation as that which was performed on A to obtain
B. Conversely, if E is an elementary m × m [n × n] matrix, then EA [AE] is the matrix
obtained from A by performing the same elementary row [column] operation as that which
produces E from Im [In ].

12
Linear Algebra 2.2. THE RANK OF A MATRIX AND MATRIX INVERSES

Theorem 5. Elementary matrices are invertible, and the inverse of an elementary matrix is
an elementary matrix of the same type.

2.2 The Rank of a Matrix and Matrix Inverses

Definition 5. If A ∈ Mm×n , the rank of A is defined to be the rank of the linear transfor-
mation LA : F n → F m , denoted by rank(A).

Theorem 6. Let A be an m × n matrix. If P and Q are invertible m × m and n × n matrices,

respectively, then
(a) rank(AQ) = rank(A),
(b) rank(P A) = rank(A),
(c) rank(P AQ) = rank(A).

Proof. (a) Observe that

R(LAQ ) = R(LA LQ ) = LA LQ (F n ) = LA (F n ) = R(LA )

since A is onto. Therefore,

rank(AQ) = dim(R(LAQ )) = dim(R(LA )) = rank(A).

(b) Since P is invertible, LP is an isomorphism. So

rank(P A) = dim(LP A ) = dim(LP LA ) = dim(LA ) = rank(A).

(c) Applying (a) and (b), we have

rank(P AQ) = rank(P A) = rank(A).

Corollary. Elementary row and column operations on a matrix are rank-preserving.

Theorem 7. The rank of any matrix equals the maximum number of its linearly independent
columns; that is, the rank of a matrix is the dimension of the subspace generated by its
columns.

Proof. For any A ∈ Mm×n (F ),

rank(A) = rank(LA ) = dim(R(LA )).

CHAPTER 2. ELEMENTARY MATRIX OPERATIONS AND SYSTEMS OF LINEAR 13

EQUATIONS
Linear Algebra 2.3. SYSTEMS OF LINEAR EQUATIONS - THEORETICAL ASPECTS

Let β be the standard ordered basis for F n . Then β spans F n and hence

R(LA ) = span(LA (β)) = span({LA (e1 ), LA (e2 ), . . . , LA (en )}).

But, for any j, we have LA (ej ) = Aej = aj , where aj is the jth column of A. Hence

R(LA ) = span({a1 , a2 , . . . , an }).

Thus
rank(A) = dim(R(LA )) = dim(span({a1 , a2 , . . . , an })).

2.3 Systems of Linear Equations - Theoretical Aspects

Definition 6. A system Ax = b of m linear equations in n unknowns is said to be homoge-
neous if b = 0. Otherwise the system is said to be nonhomogeneous.

Theorem 8. Let Ax = 0 be a homogeneous system of m linear equations in n unknowns

over a field F . Let K denote the set of all solutions to Ax = 0, Then K = N (LA ); hence K
is a subspace of F n of dimension n − rank(LA ) = n − rank(A).

Corollary. If m < n, the system Ax = 0 has a nonzero solution.

Proof. Suppose that m < n. Then rank(A) = rank(LA ) ≤ m. Hence

dim(K) = n − rank(LA ) ≥ n − m ≥ 0,

6 {0}. Thus there exists a nonzero vector s ∈ K;

where K = N (LA ). Since dim(K) > 0, K =
so s is a nonzero solution to Ax = 0.

Theorem 9. Let K be the solution set of a system of linear equations Ax = b, and let KH
be the solution set of the corresponding homogeneous system Ax = 0. Then for any solution
s to Ax = b
K = {s} + KH = {s + k | k ∈ KH }.

Proof. Let s be any solution to Ax = b. If w ∈ K, then Aw = b. Hence

A(w − s) = Aw − As = b − b = 0.

So w − s ∈ KH . Thus there exists k ∈ KH such that w − s = k. It follows that w = s + k ∈

{s} + KH , and therefore
K ⊆ {s} + KH .
Conversely, suppose that w ∈ {s} + KH ; then w = s + k for some k ∈ KH . But then

Aw = A(s + k) = As + Ak = b + 0 = b;

CHAPTER 2. ELEMENTARY MATRIX OPERATIONS AND SYSTEMS OF LINEAR 14

EQUATIONS
Linear Algebra 2.4. SYSTEMS OF LINEAR EQUATIONS - COMPUTATIONAL ASPECTS

so w ∈ K. Therefore {s} + KH ⊆ K, and thus K = {s} + KH .

Theorem 10. Let Ax = b be a system of n linear system of n linear equations in n unknowns.

If A is invertible, then the system has exactly one solution, namely, A−1 b. Conversely, if the
system has exactly one solution, then A is invertible.

Proof. Suppose that A is invertible. Substituting A−1 b into the system, we have

A(A−1 b) = (AA−1 )b = b.

Thus A−1 b is a solution. If s is an arbitrary solution, then As = b. Multiplying both sides

by A−1 gives s = A−1 b. Thus the system has one and only one solution, namely, A−1 b.
Conversely, suppose that the system has exactly one solution s. Let KH denote the
solution set for the corresponding homogeneous system Ax = 0. By the previous theorem,
{s} = {s} + KH . Thus KH = {0} and therefore N (LA ) = 0. Hence A is invertible.

Theorem 11. Let Ax = b be a system of linear equations. Then the system is consistent if
and only if rank(A) = rank(A|b).

Proof. To show that Ax = b has a solution is equivalent to showing that b ∈ R(LA ). Note
that
R(LA ) = span({a1 , a2 , . . . , an })
the span of the columns of A. Thus

Ax = b has a solution
⇐⇒ b ∈ span({a1 , a2 , . . . , an })
⇐⇒ span({a1 , a2 , . . . , an }) = span({a1 , a2 , . . . , an , b})
⇐⇒ dim(span({a1 , a2 , . . . , an })) = dim(span({a1 , a2 , . . . , an , b}))
⇐⇒ rank(A) = rank(A|b).

2.4 Systems of Linear Equations - Computational Aspects

Definition 7. Two systems of linear equations are called equivalent if the have the same
solution set.

Theorem 12. Let Ax = b be a system of m linear equations in n unknowns, and let C be an

invertible m × m matrix. Then the system (CA)x = Cb is equivalent Ax = b.

Corollary. Let Ax = b be a system of m linear equations in n unknowns. If (A0 |b0 ) is obtained

from (A|b) by a finite number of elementary row operations, then the system A0 x = b0 is
equivalent to Ax = b.

CHAPTER 2. ELEMENTARY MATRIX OPERATIONS AND SYSTEMS OF LINEAR 15

EQUATIONS
Linear Algebra 2.4. SYSTEMS OF LINEAR EQUATIONS - COMPUTATIONAL ASPECTS

Definition 8. A matrix is said to be in reduced row echelon form if the following three
conditions are satisfied.
(a) Any row containing a nonzero entry precedes any row in which all the entries are zero
(if any).

(b) The first nonzero entry in each row is the only nonzero entry in its column.
(c) The first nonzero entry in each row is 1 and it occurs in a column to the right of the
first nonzero entry in the preceding row.

CHAPTER 2. ELEMENTARY MATRIX OPERATIONS AND SYSTEMS OF LINEAR 16

EQUATIONS
Chapter 3

Determinants

3.1 Determinant of Order 2

Definition 9. If  
a b
A=
c d
is a 2 × 2 matrix with entries from a field F , then we define the determinant of A, denoted
by det(A) or |A|, to be the scalar ad − bc.

Example. For the matrices

   
1 1 1 0
A= and B =
0 1 1 1

in M2×2 (R), we have

det(A) = 1 · 1 − 1 · 0 = 1,
det(B) = 1 · 1 − 0 · 1 = 1,

and  
2 1
det(A + B) = det = 2 · 2 − 1 · 1 = 1 6= det(A) + det(B).
1 2

Theorem 13. The function det : M2×2 (F ) → F is a linear function of each row of a 2 × 2
matrix when the other row is held fixed. That is, if u, v and w are in F 2 and k is a scalar,
then      
u + kv u v
det = det + k det
w w w
and      
w w w
det = det + k det .
u + kv u v

17
Linear Algebra 3.1. DETERMINANT OF ORDER 2

Proof. Let u = (a1 , a2 ), v = (b1 , b2 ), and c = (c1 , c2 ) be in F 2 and k be a scalar. Then

       
u v a1 a2 b1 b2
det + k det = det + k det
w w c1 c2 c1 c2
= (a1 c2 − a2 c1 ) + k(b1 c2 − b2 c1 )
= (a1 + kb1 )c2 − (a2 + kb2 )c1
 
a + kb1 a2 + kb2
= det 1
c1 c2
 
u + kv
= det
w

A similar calculation shows that

     
w w w
det + k det = det .
u v u + kv

Theorem 14. Let A ∈ M2×2 (F ). Then det(A) 6= 0 if and only if A is invertible.

Proof. Suppose that det(A) 6= 0. We define a matrix

 
1 A22 −A12
M= .
det(A) −A21 A11

A straightforward calculation shows that AM = M A = I, and so A is invertible.

Conversely, suppose that A is invertible. The rank of
 
A11 A12
A=
A21 A22

must be 2. Hence A11 6= 0 or A21 6= 0. If A11 6= 0, add −A21 /A11 times row 1 of A to row 2
to obtain the matrix  
A11 A12
 A12 A21  .
0 A22 −
A11
Because elementary row operations are rank-preserving, it follows that
A12 A21
A22 − 6= 0.
A11

Therefore det(A) = A11 A22 − A12 A21 6= 0. On the other hand, if A21 6= 0, we see that
det(A) 6= 0 by adding −A11 /A21 times row 2 of A to row 1 and applying a similar argument.
Thus, in either case, det(A) 6= 0.

CHAPTER 3. DETERMINANTS 18
Linear Algebra 3.1. DETERMINANT OF ORDER 2

Theorem 15. The classical adjoint of a 2 × 2 matrix A ∈ M2×2 (F ) is the matrix

 
A22 −A12
C= .
−A12 A11

Then we have
(a) CA = AC = (det(A))I.

(b) det(C) = det(A).

(c) The classical adjoint of At is C t .
(d) If A is invertible, then
 
−1 −1 1 A22 −A12
A = (det(A)) C= .
det(A) −A21 A11

Proof. (a) A straightforward calculation shows that

  
A22 −A12 A11 A12
CA =
−A21 A11 A21 A22
   
A22 A11 − A12 A21 A22 A12 − A12 A22 det(A) 0
= =
−A21 A11 + A11 A21 −A21 A12 + A11 A22 0 det(A)

and
  
A11 A12 A22 −A12
AC =
A21 A22 −A21 A11
   
A11 A22 − A12 A21 −A11 A12 + A12 A11 det(A) 0
= = .
A21 A22 − A22 A21 −A21 A12 + A22 A11 0 det(A)

So

CA = AC = det(A)I.

(b) A straightforward calculation shows that

det(C) = A22 A11 − (−A12 )(−A21 ) = det(A).

(c) The classical adjoint of At is

(At )22 −(At )12

   
A22 −A21
= = C t.
−(At )21 (At )11 −A12 A11

(d) Note that if A is invertible, then det(A) 6= 0. Since

A(det(A)−1 C) = det(A)−1 (AC) = det(A)−1 (det(A)I) = I

CHAPTER 3. DETERMINANTS 19
Linear Algebra 3.2. DETERMINANT OF ORDER N

and
(det(A)−1 C)A = det(A)−1 (CA) = det(A)−1 (det(A)I) = I,
we have
A−1 = det(A)−1 C.

3.2 Determinant of Order n

Definition 10. Let A ∈ Mn×n (F ). If n = 1, so that A = (A11 ), we define det(A) = A11 . For
n ≥ 2, we define det(A) recursively as
n
X
det(A) = (−1)1+j A1j · det(Ã1j )
j=1

where Ãij denotes the matrix obtained from A by deleting row i and column j. The scalar
det(A) is called the determinant of A and is also denoted by |A|. The scalar

(−1)i+j det Ãij

is called the cofactor of the entry of A in row i, column j.

Note that, for 2 × 2 matrices, this definition of the determinant of A agrees with the previous
one because

det(A) = A11 (−1)1+1 det(Ã11 ) + A12 (−1)1+2 det(Ã12 ) = A11 A22 − A12 A21 .

Theorem 16. The determinant of an n × n matrix is a linear function of each row when the
remaining rows are held fixed. That is, for 1 ≤ r ≤ n, we have
     
a1 a1 a1
 ..   ..   .. 
 .   .   . 
     
 ar−1  ar−1  ar−1 
     
det u + kv  = det  u  + k det  v 
    
 ar+1  ar+1  ar+1 
     
 .   .   . 
 ..   ..   .. 
an an an

whenever k is a scalar and u, v, and each ai are row vectors in F n .

Proof. The proof is by mathematical induction on n. The result is immediate if n = 1.

Assume that the theorem is true for any (n − 1) × (n − 1) matrix where n ≥ 2. Let A be an
n × n matrix with rows a1 , a2 , . . . , an , respectively, and suppose that for some r (1 ≤ r ≤ n),
we have ar = u + kv for some u, v ∈ F n and some k ∈ F . Let u = (b1 , b2 , . . . , bn ) and
v = (c1 , c2 , . . . , cn ), and let B and C be the matrices obtained from A by replacing row r of

CHAPTER 3. DETERMINANTS 20
Linear Algebra 3.2. DETERMINANT OF ORDER N

A by u and v, respectively, i.e.,

 

a1 a1
 ..   .. 
 .   . 
   
ar−1  ar−1 
   
 u 
B=  v 
and C = 
 
ar+1  ar+1 
   
 .   . 
 ..   .. 
an an

We must prove that det(A) = det(B) + k det(C). For r = 1, we have

n
X
det(A) = (−1)1+j A1j · det(Ã1j )
j=1
n
X
= (−1)1+j (bj + kcj ) · det(Ã1j )
j=1
Xn n
X
= (−1)1+j bj · det(Ã1j ) + k (−1)1+j cj · det(Ã1j )
j=1 j=1
n
X n
X
= (−1)1+j B1j · det(B̃1j ) + k (−1)1+j C1j · det(C̃1j )
j=1 j=1

= det(B) + k det(C).

For r > 1 and 1 ≤ j ≤ n, the rows of Ã1j , B̃1j , and C̃1j are the same except for row r − 1.
Moreover row r − 1 of Ã1j is

(b1 + kc1 , . . . , bj−1 + kcj−1 , bj+1 + kcj+1 , . . . , bn + kcn ),

which is the sum of row r − 1 of B̃1j and k times row r − 1 of C̃1j . Since B̃1j and C̃1j are
(n − 1) × (n − 1) matrices, we have

det(Ã1j ) = det(B̃1j ) + k det(C̃1j )

by induction hypothesis. Thus since A1j = B1j = C1j , we have

n
X
det(A) = (−1)1+j A1j · det(Ã1j )
j=1
Xn
= (−1)1+j A1j · (det(B̃1j ) + k det(C̃1j ))
j=1
n
X n
X
1+j
= (−1) A1j · det(B̃1j ) + k (−1)1+j A1j · det(C̃1j )
j=1 j=1

= det(B) + k det(C).

CHAPTER 3. DETERMINANTS 21
Linear Algebra 3.2. DETERMINANT OF ORDER N

This shows that the theorem is true for n × n matrices, and so the theorem is true for all
square matrices by mathematical induction.

Corollary. If A ∈ Mn×n (F ) has a row consisting entirely of zeros, then det(A) = 0.

Proof. Let A ∈ Mn×n (F ) with rows a1 , a2 , . . . , an , respectively, and suppose that ar = 0 ∈

F n for some r (1 ≤ r ≤ n). Then ar = 0 = 0 + 0 = ar + ar and we have
       
a1 a1 a1 a1
 ..   ..   ..   .. 
.  .  . .
       
det(A) = det  0  = det 0 + 0 = det  0  + det 
     
 0  = det(A) + det(A).

.  .  . .
 ..   ..   ..   .. 
an an an an

Therefore det(A) = 0.

Lemma. Let B ∈ Mn×n (F ), where n ≥ 2. If row i of B equals ek for some k (1 ≤ k ≤ n),

then det(B) = (−1)i+k det(B̃ik ).

Proof. The proof is by mathematical induction on n. The lemma is clearly true for n = 2.
Assume that for some integer n ≥ 3, the lemma is true for (n − 1) × (n − 1) matrices, and
let B ∈ Mn×n (F ) in which row i of B equals ek for some k (1 ≤ k ≤ n). The result follows
immediately from the definition of the determinant if i = 1. Suppose that 1 < i ≤ n. For
each j 6= k (1 ≤ j ≤ n), let Cij denote the (n − 2) × (n − 2) matrix obtained from B by
deleting rows 1 and i and columns j and k. For each j, row i − 1 of B̃1j is the following
vector in F n−1 : 
ek−1 if j < k

0 if j = k

ek if j > k.


Hence by induction hypothesis, we have


(i−1)(k−1)
(−1)
 det(Cij ) if j < k
det(B̃1j ) = 0 if j = k
 (i−1)k
(−1) det(Cij ) if j > k.


CHAPTER 3. DETERMINANTS 22
Linear Algebra 3.2. DETERMINANT OF ORDER N

Therefore
n
X
det(B) = (−1)1+j B1j · det(B̃1j )
j=1
X X
= (−1)1+j B1j · det(B̃1j ) + (−1)1+j B1j · det(B̃1j )
j<k j>k
X  
= (−1)1+j B1j · (−1)(i−1)(k−1) det(Cij )
j<k
X  
+ (−1)1+j B1j · (−1)(i−1)k det(Cij )
j>k
 
X X
= (−1)i+k  (−1)1+j B1j · det(Cij ) + (−1)1+(j−1) B1j · det(Cij )
j<k j>k
i+k
= (−1) det(B̃ik ).

This shows that the lemma is true for n × n matrices, and so the lemma is true for all square
matrices by mathematical induction.

Theorem 17. The determinant of a square matrix can be evaluated by cofactor expansion
along any row. That is, if A ∈ Mn×n (F ), then for any integer i (1 ≤ i ≤ n),
n
X
det(A) = (−1)i+j Aij · det(Ãij ).
j=1

Proof. Cofactor expansion along the first row of A gives the determinant of A by definition.
So the result is true if i = 1. Fix i > 1. Row i of A can be written as
n
X
Aij ej .
j=1

For 1 ≤ j ≤ n, let Bj denote the matrix obtained from A by replacing row i of A by ej .

Then
Xn n
X
det(A) = Aij det(Bj ) = (−1)i+j Aij · det(Ãij ).
j=1 j=1

Corollary. If A ∈ Mn×n (F ) has two identical rows, then det(A) = 0.

Proof. The proof is by mathematical induction on n. The corollary is clearly true for n = 2.
Assume that for some n ≥ 3, it is true for (n − 1) × (n − 1) matrices, and let rows r and s of
A ∈ Mn×n (F ) be identical for r 6= s. Since n ≥ 3, we can choose i (1 ≤ i ≤ n) other than r
and s. Now
X n
det(A) = (−1)i+j Aij · det(Ãij )
j=1

CHAPTER 3. DETERMINANTS 23
Linear Algebra 3.2. DETERMINANT OF ORDER N

Since each Ãij is an (n − 1) × (n − 1) matrix with two identical rows, the induction hypothesis
implies that each det(Ãij ) = 0, and hence det A = 0. This completes the proof for n × n
matrices, and so the corollary is true for all square matrices by mathematical induction.

Theorem 18. If A ∈ Mn×n (F ) and B is a matrix obtained from A by interchanging any two
rows of A, then det(B) = − det(A).

Proof. Let the rows of A be a1 , a2 , . . . , an , and let B be the matrix obtained from A by
interchanging rows r and s, where r < s. Thus
   
a1 a1
 ..   .. 
. .
   
 ar   as 
   
A =  .  and B =  ...  .
 ..   
   
 as   ar 
   
. .
 ..   .. 
an an

Consider the matrix obtained from A by replacing rows r and s by ar + as . Then

     
a1 a1 a1
 ..   ..   .. 
 .   .   . 
     
ar + as   ar   as 
     
0 = det  ...  = det  ...  + det  ... 
     
     
ar + as  ar + as  ar + as 
     
 .   .   . 
 ..   ..   .. 
an an an
       
a1 a1 a1 a1
 ..   ..   ..   .. 
. . . .
       
 ar   ar   as   as 
       
= det  .  + det  .  + det  .  + det  ... 
 ..   ..   ..   
       
 ar   as   ar   as 
       
. . . .
 ..   ..   ..   .. 
an an an an
= 0 + det(A) + det(B) + 0.

Therefore det(B) = − det(A).

Theorem 19. Let A ∈ Mn×n (F ), and let B be a matrix obtained by adding a multiple of
one row of A to another row of A. Then det(A) = det(B).

CHAPTER 3. DETERMINANTS 24
Linear Algebra 3.3. PROPERTIES OF DETERMINANTS

Proof. Suppose that B is the n × n matrix obtained from A by adding k times row r to
row s, where r 6= s. Let the rows of A be a1 , a2 , . . . , an , and the rows of B be b1 , b2 , . . . , bn .
6 s and bs = as + kar . Let C be the matrix obtained from A by replacing
Then bi = ai for i =
row s with ar . Then det(C) = 0 and we have

det(B) = det(A) + k det(C) = det(A).

Corollary. If A ∈ Mn×n (F ) has rank less than n, then det(A) = 0.

Proof. Since rank(A) ≤ n, the rows a1 , a2 , . . . , an of A are linearly dependent. Then some
row of A, say, row r, is a linear combination of the other rows. So there exist ci s.t.

ar = c1 a1 + · · · + cr−1 ar−1 + cr+1 ar+1 + · · · + cn an .

Let B be the matrix obtained from A by adding −ci times row i to row r for each i 6= r.
Then row r of B consists entirely of zeros, and so det(B) = 0. Hence, by the previous
theorem,
det(A) = det(B) = 0.

Remark. The following rules summarize the effect of an elementary row operation on the
determinant of a matrix A ∈ Mn×n (F ).

1. If B is a matrix obtained by interchanging any two rows of A, then det(B) = − det(A).

2. If B is a matrix obtained by multiplying a row of A by a nonzero scalar k, then
det(B) = k det(A).
3. If B is a matrix obtained by adding a multiple of one row of A to another row of A,
then det(B) = det(A).

3.3 Properties of Determinants

Proposition 1. The following are the facts about the determinants of elementary matrices:
Let E be an elementary matrix.
1. If E is obtained by interchanging any two rows of I, then det(E) = −1.
2. if E is obtained by multiplying a row of I by a nonzero scalar k, then det(E) = k.
3. If E is obtained by adding a multiple of one row of I to another row, then det(E) = 1.

4. det(E t ) = det(E).

Theorem 20. For any A, B ∈ Mn×n (F ), det(AB) = det(A) · det(B).

Proof. If A is an elementary matrix, then the theorem is clearly true.

CHAPTER 3. DETERMINANTS 25
Linear Algebra 3.3. PROPERTIES OF DETERMINANTS

If A is an n × n matrix with rank less than n, then det(A) = 0. Since rank(AB) ≤

rank(A) < n, we have det(AB) = 0. Thus det(AB) = det(A) det(B) in this case.
On the other hand, if A has rank n, then A is invertible and hence is a product of
elementary matrices, say, A = Em · · · E2 E1 . Then

det(AB) = det(Em · · · E2 E1 B)
= det(Em ) det(Em−1 · · · E1 B)
..
.
= det(Em ) · · · det(E2 ) det(E1 ) det(B)
= det(Em · · · E2 E1 ) det(B)
= det(A) det(B)

Corollary. A matrix A ∈ Mn×n (F ) is invertible if and only if det(A) 6= 0. Moreover, if

1
A ∈ Mn×n (F ) is invertible, then det(A−1 ) = .
det(A)

Proof. If A ∈ Mn×n (F ) is not invertible, then rank(A) < n and therefore det(A) = 0.
Conversely, if A ∈ Mn×n (F ) is invertible, then

det(A) det(A−1 ) = det(AA−1 ) = det(I) = 1.

1
Hence det(A) 6= 0 and det(A−1 ) = .
det(A)

Theorem 21. For any A ∈ Mn×n (F ), det(At ) = det(A).

Proof. If rank(A) < n, then rank(At ) = rank(A) < n and therefore det(At ) = 0 = det(A).
On the other hand, if rank(A) = n, then A is a product of elementary matrices, say,
A = Em · · · E2 E1 . Since det(Eit ) = det(Ei ) for every i, we have

det(At ) = det(E1t E2t · · · Em

t
)
= det(E1t ) det(E2t ) · · · det(Em
t
)
= det(E1 ) det(E2 ) · · · det(Em )
= det(Em ) · · · det(E2 ) det(E1 )
= det(Em · · · E2 E1 )
= det(A).

Proposition 2. The determinant of an n × n upper triangular matrix is the product of its

diagonal entries.

Proof. The proof is by mathematical induction on n. The proposition is clearly true for
n = 1. Assume that for some n ≥ 2, it is true for (n − 1) × (n − 1) matrices. Since Ãnn is

CHAPTER 3. DETERMINANTS 26
Linear Algebra 3.3. PROPERTIES OF DETERMINANTS

an (n − 1) × (n − 1) upper triangle matrix, the induction hypothesis implies that

n−1
Y
det(Ãnn ) = Aii .
i=1

Evaluating det(A) by cofactor expansion along row n gives

n
X
det(A) = (−1)n+j Anj · det(Ãnj )
j=1

= (−1)2n Ann · det(Ãnj )

n
Y
= An n · Aii
i=1
n
Y
= Aii .
i=1

This completes the proof for n × n matrices, and so the corollary is true for all upper triangle
matrices by mathematical induction.

Lemma. If M ∈ Mn×n (F ) can be written in the form

 
A O
M= ,
O I

where A is a square matrix, then det(M ) = det(A).

Proof. Suppose that A ∈ Mm×m (F ) for some m (1 ≤ m < n). For each k (1 ≤ k ≤ m − n),
let Mk be the matrix in the form  
A O
.
O Ik
The proof is by mathematical induction on k. For k = 1, the determinant of M1

det(M1 ) = 0 + · · · + 0 + (−1)(m+1)+(m+1) 1 · det((M

f1 )(m+1)(m+1) ) = det(A)

by cofactor expansion along row m + 1. Suppose that for some k ≥ 2, the lemma is true for
Mk−1 , i.e., det(Mk−1 ) = det(A). Then

det(Mk ) = 0 + · · · + 0 + (−1)(m+k)+(m+k) 1 · det((M

gk )(m+k)(m+k) )
= det(Mk−1 ) = det(A).

by cofactor expansion along row m + k. So the lemma holds for Mn and thus for all Mk by
mathematical induction. Since A is arbitrary square matrix, this completes the proof.

CHAPTER 3. DETERMINANTS 27
Linear Algebra 3.3. PROPERTIES OF DETERMINANTS

Lemma. If M ∈ Mn×n (F ) can be written in the form

 
I O
M= ,
O A

where A is a square matrix, then det(M ) = det(A).

Proof. It is trivial by mathematical induction.

Proposition 3. If M ∈ Mn×n (F ) can be written in the form

 
A B
M= ,
O C

where A and C are square matrices, then det(M ) = det(A) det(C).

Proof. Note that M can be written as

   
I O I B A O
M= .
O C O I O I

Then, by the lemma,

   
I O I B A O
det(M ) = det
O C O I O I
     
I O I B A O
= det det det
O C O I O I
= det(C) · 1 · det(A) = det(A) det(C)

Theorem 22 (Cramer’s Rule). Let Ax = b be the matrix form of a system of n linear equations
in n unknowns, where x = (x1 , x2 , . . . , xn )t . If det(A) 6= 0, then this system has a unique
solution, and for each k (1 ≤ k ≤ n),

det(Mk )
xk = ,
det(A)

where Mk is the n × n matrix obtained from A by replacing column k of A by b.

Proof. If det(A) 6= 0, then A is invertible and therefore Ax = b has a unique solution. For
each k (1 ≤ k ≤ n), let ak denote the kth column of A and Xk denote the matrix obtained
from the In by replacing column k by x. Then AXk is the n × n matrix whose ith column is

Aei = ai if i 6= k and Ax = b if i = k.

Thus AXk = Mk . Evaluating det(Xk ) by cofactor expansion along row k gives

det(Xk ) = xk · det(In−1 ) = xk .

CHAPTER 3. DETERMINANTS 28
Linear Algebra 3.3. PROPERTIES OF DETERMINANTS

Hence
det(Mk ) = det(AXk ) = det(A) det(Xk ) = det(A)xk .
Therefore
det(Mk )
xk = .
det(A)

Example. Let T : Pn (F ) → F n+1 be the linear transformation defined by

T (f ) = (f (x0 ), f (x1 ), . . . , f (xn )),

where x0 , x1 , . . . , xn are distinct scalars in an infinite field F . Let β be the standard ordered
basis for Pn (F ) and γ be the standard ordered basis for F n+1 . Then

1 x0 x20 · · · xn0
 
1 x1 x21 · · · xn1 
[T ]γβ =  . ..  .
 
.. ..
 .. . . . 
1 xn x2n ··· xnn

A matrix with this form is called a Vandermonde matrix. First, Apply elementary column
operation of type 3 by subtracting to each column, except the first one, the preceding column
multiplied by x1 . This gives the matrix
 
1 0 0 ··· 0
1 x1 − x0 x1 (x1 − x0 ) · · · xn−1 1 (x1 − x0 ) 
1 x2 − x0 x2 (x2 − x0 ) · · · xn−1 (x2 − x0 ) 
 
 2 .
. .. .. ..
 ..

. . . 
n−1
1 xn − x0 xn (xn − x0 ) · · · xn (xn − x0 )

CHAPTER 3. DETERMINANTS 29
Linear Algebra 3.3. PROPERTIES OF DETERMINANTS

Evaluating the determinant by cofactor expansion along the first row gives

x1 − x0 x1 (x1 − x0 ) · · · xn−1
 
1 (x1 − x0 )
 x2 − x0 x2 (x2 − x0 ) · · · xn−1 (x2 − x0 ) 
2
det([T ]γβ ) = det 
 
.. .. .. 
 . . . 
xn − x0 xn (xn − x0 ) · · · xnn−1 (xn − x0 )
1 x1 x21 · · · xn−1
 
1
n 1 x2 x22 · · · x n−1 
Y 2
= (xi − x0 ) det  .
 
. .
. .
. .
.

i=1
. . . . 
1 xn x2n · · · xn−1 n
1 x2 x22 · · · xn−2
 
2
n n 1 x3 x23 · · · xn−2 
Y Y 3
= (xi − x0 ) (xi − x1 ) det  .
 
.. .. .. 
i=1 i=2
 .. . . . 
1 xn x2n ··· xn−2
n
..
.
Y
= (xi − xj ).
0≤j<i≤n

Exercise. Let A ∈ Mn×n (F ) have the form

 
0 0 0 ··· 0 a0
−1 0 0 ··· 0 a1 
 
A= 0
 −1 0 ··· 0 a2 

 .. .. .. .. .. 
 . . . . . 
0 0 0 ··· −1 an−1

Compute det(A + tIn ).

Proof. Claim that

det(A + tIn ) = a0 + a1 t + a2 t2 + · · · + an−1 tn−1 + tn .

The proof is by mathematical induction on n. For n = 1, it is obviously that

det( a0 + tI1 ) = a0 + t.

Suppose that for some n ≥ 2, the statement holds for the case n − 1. By the induction
hypothesis,
 
t 0 ··· 0 a1
−1 t · · · 0 a2 
det  .  = a1 + a2 t + · · · + an−1 tn−2 + tn−1 .
 
.. .. ..
 .. . . . 
0 0 ··· −1 an−1 + t

CHAPTER 3. DETERMINANTS 30
Linear Algebra 3.3. PROPERTIES OF DETERMINANTS

Then
 
t 0 0 ··· 0 a0
−1 t 0 ··· 0 a1 
 
det  0
 −1 t ··· 0 a2 

 .. .. .. .. .. 
 . . . . . 
0 0 0 · · · −1 an−1 + t
 
  −1 t 0 ··· 0
t 0 ··· 0 a1
−1
0 −1 t ··· 0 
t ··· 0 a2   
= t det  .

.. .. ..  + (−1)1+n a0 det  0
  0 −1 ··· 0 
 ..

. . .   .. .. .. ..
 . . . .
0 0 · · · −1 an−1 + t
0 0 0 ··· −1
n−2 n−1 1+n n−1
= t(a1 + a2 t + · · · + an−1 t +t ) + (−1) a0 (−1)
n−1 n
= a0 + a1 t + · · · + an−1 t +t .

Hence the statement holds for all square matrices in the form by mathematical induction.

Definition 11. The classical adjoint of a square matrix A is the transpose of the matrix
whose ij-entry is the ij-cofactor of A.

Proposition 4. Let C be the classical adjoint of a matrix A ∈ Mn×n (F ). Then

AC = (det(A))I˙n .

Proof. Fix j (1 ≤ j ≤ n). Consider the system Ax = ej . For each k, let ak denote the kth
column of A and Xk denote the matrix obtained from In by replacing column k by x. Then
AXk is the matrix whose the ith column is

Aei = ai if i 6= k and Ax = ej if i = k.

Thus AXk = Mk , where Mk is the matrix obtained from A by replacing column k by ej .

Then
det(AXk ) = det(A) det(Xk ) = det(A)xk
and
det(Mk ) = (−1)j+k det Ãjk = Ckj .
So
Ckj = det(A)xk ;
it follows that  
C1j
 C2j 
A  .  = A(det(A) · xk ) = det(A) · ej .
 
.
 . 
Cnj

CHAPTER 3. DETERMINANTS 31
Linear Algebra 3.3. PROPERTIES OF DETERMINANTS

Hence
AC = det(A) · In .

Proof. Fix k (1 ≤ k ≤ n). Evaluating the determinant of A by cofactor expansion along row
k gives
Xn n
X
det(A) = (−1)k+j Akj · det(Ãkj ) = Akj Cjk .
j=1 j=1

For each i 6= k (1 ≤ i ≤ n), let Bi denote the matrix obtained from A by replacing row k by
row i. Then det(Bi ) = 0 since Bi has two identical row and the kj-cofactors of A and Bi
are the same for each j. It follows that
n
X n
X
Aij Cjk = (Bi )kj Cjk = det(Bi ) = 0.
j=1 j=1

Thus, for each k,  

C1k
 C2k 
A  .  = det(A) · ek .
 
 .. 
Cnk
So we have
AC = det(A) · I.

Corollary. Let C be the classical adjoint of a matrix A ∈ Mn×n (F ). Then the following hold:
1. If A is invertible, then
1
A−1 = · C.
det(A)

2. det(C) = det(A)n−1 .

CHAPTER 3. DETERMINANTS 32
Index

Annihilator, 10 Equivalent, 15

Classical adjoint, 19, 31 Homogeneous, 14

Cofactor, 20

Determinant, 17, 20 Linear functional, 3

Dual basis, 4
Dual space, 3 Rank, 13
Reduced row echelon form, 16
Elementary matrix, 12
Elementary operation, 12 Vandermonde matrix, 29

33