We now introduce a new algebraic structure the matrix.

. Matrices are twodimensional arrays created by arranging vectors into rows and columns. We
examine several fundamental types of matrices, as well as three basic operations on matrices and their properties.
Definition of a Matrix
DEFINITION

An m x n matrix is a rectangular array of real numbers, arranged in m

rows and n columns. The elements of a matrix are called the entries. The
expression m x n denotes the size of the matrix.
For example, each of the following is a matrix, listed with its correct size:
4 -2
B =

1 7

-5
matrix

C =

1 2 3
4 56
7 8 9
3 x 3 matrix

3 x 2 matrix

D =

1
- 2
3 x 1 matrix

E = [ 4 - 3 0]
1 x 3 matrix

1 x 1 matrix

Here are some conventions to remember regarding matrices.

 An m x n matrix can be thought of either as a collection of m row vectors,

each having n coordinates, or as a collection of n column vectors, each
having in coordinates. A matrix with just one row (or column) is essentially equivalent to a vector with coordinates in row (or column) form.
We often write a, j to represent the entry in the /th row and jth column
of a matrix A. For example, in the previous matrix A, azi is the entry 5
in the second row and third column. A typical 3 x 4 matrix C has entries
symbolized by
C =

Cll C12 C13 C14

('21 c22 c23 C'24
C31 C-.i2 t'33 C;)4

Mm represents the set of all matrices with real-number entries having

111 rows and n columns. For example, M34 is the set of all matrices having
three rows and four columns. A typical matrix in Mm has the form of
the preceding matrix C.

The main diagonal entries of a matrix A are o n , 1122. f33

, those
that lie on a diagonal line drawn down to the right, beginning from the
upper-left corner of the matrix.

Two 111 x 7? matrices A and B are equal if and only if all of their corresponding
entries are equal. That is, A = B if a,j = bjj for all /, 1 < i < in, and for all
J- 1 < j

< n.

Note that the following may be considered equal as vectors but not as
matrices:
3
- 2
[3, - 2 . 4] and
4
since the former is a 1 x 3 matrix, but the latter is a 3 x 1 matrix.
Special Types of Matrices
We now describe a few important types of matrices.
A square matrix is an n x n matrix; that is, a matrix having the same number
of rows as columns. For example, the following matrices are square:
5 0
A 9 -2

and

1 2 3'
4 56
7 8 9

B -

A diagonal matrix is a square matrix in which all entries that are not on
the main diagonal are zero. That is, A is diagonal if and only if
== 0 for / ^ j.
For example, the following are diagonal matrices:

6 0
0 7
0

0
0

F =

0 - 2

0
II
0
0

4 0
0
0 0
0
0 0 -2
0 0
0

and

-[1!]

However, the matrices are not diagonal. (The main diagonal elements have
been shaded in each case.) We use V,, to represent the set of all n x n diagonal
matrices.

-[

and

J =

"

0
7
5

4
0

3'

-2

An identity matrix is a diagonal matrix with all main diagonal entries equal
to 1. That is, an n x n matrix A is an identity matrix if and only if cijj = 0 for
i / j and an = 1 for 1 < i < n. The n x n identity matrix is denoted by I. For
example, the following are identity matrices:
Io =

1
0

0
1

and

I4 =

1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1

If the size of the identity matrix is clear from the context, I alone may be used.
An upper triangular matrix is a square matrix with all entries below the
main diagonal equal to zero. That is, an /; x /; matrix A is upper triangular if and only if
= 0 for i > j. For example, the following are upper
triangular:

"6
p =
0
0

9
-2
0

11 "
3
5

and

R=

3
9
14

-2

-6

-2

-4
0
0

9
0
0

5
8
3

Similarly, a lower triangular matrix is one in which all entries above the
main diagonal equal zero; for example,
We use Un to represent the set of all n x n upper triangular matrices and
Cn to represent the set of all n x n lower triangular matrices.
A zero matrix is any matrix all of whose entries are zero. Om denotes the
m x n zero matrix, and O,, denotes the n x n zero matrix. For example,
0
0

0 2 :i =

0
0

0
0

and

O-) =

0
0

0
0

If the size of the zero matrix is clear from the context, O alone may two used.
Addition and Scalar Multiplication with Matrices
DEFINITION
Let A and B both be m x n matrices. The sum of A and B is the m x /;
matrix (A + B) whose (/', j) entry is equal to a,, + bij.
As wit h vectors, matrices are summed simply by adding t heir corresponding
entries together. For example,
6-321
o

'

5 - 6 -3"
-4 -2 -4

11 - 9 - 1
-11 -2
0

Notice that the definition does not allow addition of matrices with different
sizes. For example, the following matrices cannot be added:
A =

-2 3 0
1 4 -5

and

B=

6
-2
I

7
5
I

since A is a 2 x 3 matrix, and B is a 3 x 2 matrix.

DEFINITION
Let A be an m x n matrix, and let c be a scalar. Then the matrix cA. the
scalar multiplication of c and A, is the m x n matrix whose (/. /') entry
is equal to cctjj.

As with vectors, scalar multiplication with matrices is done by multiplying

every entry by the given scalar. For example, ii c
2 and
A=

4-16
7 , then 2A = - 8
- 4
2 4 9 -5

2-12-14
- 8

-18

10

Note that if A is any m x n matrix, then OA 0 m .

Let - A denote the matrix - 1 A . the scalar multiple of A by ( - 1 ) . For example, if
-3
3 -2
then 1A = A = - 1 0
A=
10 6
Also, we define subtraction of matrices as A B = A + ( - B ) .
As with vect ors, sums of scalar multiples of matrices are called linear combinations. For example, 2A+6B 3C is a linear combination ot A,B, andC.
Fundamental Properties of Addition and Scalar Multiplication
The properties in the next theorem are similar to the vector properties of

THEOREM 1.11
Let A, B. and C be m x n matrices (elements of Mmn), and let c and d be
scalars. Then
(1)
(2)
(3)

A+ B= B+ A
A + (B -f- C) = (A + B) + C
Omn + A = A + 0, = A

(4)

A + ( - A ) = ( - A ) + A = Omn

(5)
(6)
(7)

c (A -f- B) = cA -|- cB
(c + d) A = cA + dA
(cd) A = c(dA)

(8)

1(A) = A

Commutative Law of Addition

Associative Law of Addit ion
Existence of Identity Element for
Addition
Existence of Inverse Elements for
Addition
Distributive Laws of Scalar
Multiplication over Addition
Associativity of Scalar
Multiplication
Identity Property for Scalar
Multiplication

lo prove each property, calculate corresponding entries on both sides and

show they agree by applying an appropriate law of real numbers. We prove
part (1) as an example and leave some of the remaining proofs as Exercise 10.
PROOF OF THEOREM 1.11, PART ( 1 ) :
For any /. j, where 1 < i < m and 1 < j < n, the (/. j) entry of (A + B) is
the sum of the entries o,j and b, j from A and B, respectively. Similarly, the

(/, j) entry of B + A is the sum of bjj and a,y. But a, j + b,, = bij + a,/, by the
commutative property of addition for real numbers. Hence, A + B = B + A,
because their corresponding entries agree.

The Transpose of a Matrix and Its Properties

DEFINITION
If A is an m x n matrix, then the transpose, A 7 , is the n x m matrix
whose (/, j) entry is the same as the
/) entry of A.
Thus, transposing A moves the (/, j) entry of A to the ( j , /') entry of A'
Notice that the entries on the main diagonal do not move as we convert A to
A 7 . However, all entries above the main diagonal are moved below it, and vice
versa. For example,

if A =

then A

6
-2
3
1

10
4
0
8

6-231
10 4 0 8

and

B=

and

1 5-3
0-4
6
0 0-5

B7 =

1 0 0
5 -4 0
-3 6 -5

Notice that the transpose changes the rows of A into the columns of A ' .
Similarly, the columns of A become the rows of A ' . Also note that the transpose of an upper triangular matrix (such as B) is lower triangular, and vice
versa.
Three useful properties of the transpose are given in the next theorem. We
prove one and leave the others as Exercise 11.
THEOREM 1.12
Let A and B both be m x n matrices, and let c be a scalar. Then
(1)
(2)

(3)

(A7y = A
(A + B)7" = A 7 + B7"
(cA)7" = c(A 7 )

PROOF OF THEOREM 1.12, PART (2):

Notice that both (A + B) 7 and ( A ' ) + ( B ' ) are n x m matrices (why?). We need
to show that the (/', j) entries of both are equal, for 1 < i < n and 1 < j < m.
Now, the (/, j) entry of (A + B)7 equals the ( j , i) entry of A + B, which is
ciji + bjj. But the (/, j) entry of A 7 + B 7 equals the (/, j) entry of A 7 plus the
(/, j) entiy of B7', which is also cijj + bj,.

Symmetric and Skew-Symmetric Matrices

DEFINITION
A matrix A is symmetric if and only if A = A 7 . A matrix A is
skew-symmetric if and only if A = -A1.
In Exercise 5, yon are asked to show that any symmetric or skew-symmetric
matrix is a square matrix.
EXAMPLE 1

Consider the following matrices:

2
A=

4'

6 - 1

and

B=

0-3

0 - 1 3 6
1 0
2-5
-3-2 0 4
-4 0
- 6

A is symmetric and B is skew-symmetric, because their respective transposes

are
0 1-3-6
2 6 4
-1 0-2 5
A' = G - 1 0
and B' =
3 2 0 -4
4 0-3
6 - 5 4 0
which equal A and B, respectively. However, neither of the following is symmetric or skew-symmetric (why?):
C=

3-2
2 4
-1

1
0

0 - 2

1 -2

and

D=

3 4
5 -6

Notice that an n x n matrix A is symmetric [skew-symmetric] if and only

if cijj = ajj [ciij ci/i ] for all i,j such that 1 < i,j < n. In other words,
the entries above the main diagonal are reflected into equal (for symmetric)
or opposite (for skew-symmetric) entries below the diagonal. Since the main
diagonal elements are reflected into themselves, all of the main diagonal
elements of a skew-symmetric matrix must be zeroes (,, = an only if
an = 0).
Notice that any diagonal matrix is equal to its transpose, and so such matrices are automatically symmetric. Another useful result is the following:
THEOREM 1.13
Every square matrix A can be decomposed uniquely as the sum of two
matrices S and V, where S is symmetric and V is skew-symmetric.
An outline ol the proof of Theorem 1.13 is given in Exercise 13, which also
states that S =g (A + A7') and V = | (A - A7").

EXAMPLE 2

We can decompose the matrix

A=

-4 2 5'
6 37
-10 2

as the sum of a symmetric matrix S and a skew-symmetric matrix V, where

s =

~-4 2 5"
"-4 6
6 3 7 +
2 3
2 I -1
5 7
0 2

2 V

- f

(A + A 7 ) = \ (

" - 4 4 2"
4 3 2i
2 27 2_

and

/ " - 4 2 5"

6 3 7
0 2

- 1

"-4 6
2 3
5 7

- f

0
2

-2 3'
2
0 1
" 3 -I 0

Notice that S and V really are, respectively, symmetric and skew-symmetric

and that S + V really does equal A.
1

DEFINITION
I f A is a square matrix, then the trace of A, denoted by tr(A), is defined to be the sum
o f the entries on the main diagonal o f A. The trace o f A is undefined i f A is not a
square matrix.

E X A M P L E 12 Trace of a Matrix
The following are examples o f matrices and their traces.
p.
a ii
-4 =

f 13

2I

a 22 CI2 3

3I

"32

f 33

tr(A) = flu + CI22 + 033

-8

-3

_2

-1

1112

B =

tr(Z?) = - l + 5 + 7 + 0 = l l