Nullspace

Let A = (aij ) be an m×n matrix.

Definition. The nullspace of the matrix A,
denoted N(A), is the set of all n-dimensional
column vectors x such that Ax = 0.
  x   
1
a11 a12 a13 . . . a1n 0
x2   
a
 21 a22 a23 . . . a2n  
  0
 .. .. .. .. ..   x  = .
 . . . . . 
3 .
 ...   . 

am1 am2 am3 . . . amn 0
xn
The nullspace N(A) is the solution set of a system
of linear homogeneous equations (with A as the
coefficient matrix).
Let A = (aij ) be an m×n matrix.
Theorem The nullspace N(A) is a subspace of the
vector space Rn .
Proof: We have to show that N(A) is nonempty, closed under
addition, and closed under scaling.
First of all, A0 = 0 =⇒ 0 ∈ N(A) =⇒ N(A) is not empty.
Secondly, if x, y ∈ N(A), i.e., if Ax = Ay = 0, then
A(x + y) = Ax + Ay = 0 + 0 = 0 =⇒ x+y ∈ N(A).
Thirdly, if x ∈ N(A), i.e., if Ax = 0, then for any r ∈ R one
has A(r x) = r (Ax) = r 0 = 0 =⇒ r x ∈ N(A).

Definition. The dimension of the nullspace N(A) is

called the nullity of the matrix A.
Problem. Find the nullity of the matrix
 
1 1 1 1
A= .
2 3 4 5
Elementary row operations do not change the nullspace.
Let us convert A to reduced row echelon form:
     
1 1 1 1 1 1 1 1 1 0 −1 −2
→ →
2 3 4 5 0 1 2 3 0 1 2 3
 
x1 − x3 − 2x4 = 0 x1 = x3 + 2x4
⇐⇒
x2 + 2x3 + 3x4 = 0 x2 = −2x3 − 3x4
General element of N(A):
(x1 , x2 , x3 , x4 ) = (t + 2s, −2t − 3s, t, s)
= t(1, −2, 1, 0) + s(2, −3, 0, 1), t, s ∈ R.
Vectors (1, −2, 1, 0) and (2, −3, 0, 1) form a basis for N(A).
Thus the nullity of the matrix A is 2.
Row space
Definition. The row space of an m×n matrix A is
the subspace of Rn spanned by rows of A.
The dimension of the row space is called the rank
of the matrix A.
Theorem 1 Elementary row operations do not change the
row space of a matrix.
Theorem 2 If a matrix A is in row echelon form, then the
nonzero rows of A are linearly independent.
Corollary The rank of a matrix is equal to the number of
nonzero rows in its row echelon form.
Theorem 3 The rank of a matrix A plus the nullity of A
equals the number of columns of A.
Problem. Find the
 rank of the matrix

−1 0 −1 2
A =  2 0 2 0 .
1 0 1 −1
Elementary row operations do not change the row
space. Let us convert A to row echelon form:
   
−1 0 −1 2 −1 0 −1 2
 2 0 2 0 →  0 0 0 4
1 0 1 −1 1 0 1 −1
   
−1 0 −1 2 −1 0 −1 2
→  0 0 0 4 →  0 0 0 0
0 0 0 1 0 0 0 1
    
−1 0 −1 2 1 0 1 −2
→  0 0 0 1 → 0 0 0 1
0 0 0 0 0 0 0 0
Vectors (1, 0, 1, −2) and (0, 0, 0, 1) form a basis for
the row space of A. Thus the rank of A is 2.
Remark. The rank of A equals the number of
nonzero rows in the row echelon form, which equals
the number of leading entries.
The nullity of A equals the number of free variables
in the corresponding system, which equals the
number of columns without leading entries.
Consequently, rank+nullity is the number of all
columns in the matrix A.
Theorem 1 Elementary row operations do not
change the row space of a matrix.
Proof: Suppose that A and B are m×n matrices such that B
is obtained from A by an elementary row operation. Let
a1 , . . . , am be the rows of A and b1 , . . . , bm be the rows of B.
We have to show that Span(a1 , . . . , am ) = Span(b1 , . . . , bm ).
Observe that any row bi of B belongs to Span(a1 , . . . , am ).
Indeed, either bi = aj for some 1 ≤ j ≤ m, or bi = r ai for
some scalar r 6= 0, or bi = ai +r aj for some j 6= i and r ∈ R.
It follows that Span(b1 , . . . , bm ) ⊂ Span(a1 , . . . , am ).
Now the matrix A can also be obtained from B by an
elementary row operation. By the above,
Span(a1 , . . . , am ) ⊂ Span(b1 , . . . , bm ).
Problem. Find the nullity of the matrix
 
1 1 1 1
A= .
2 3 4 5

Alternative solution: Clearly, the rows of A are

linearly independent. Therefore the rank of A is 2.
Since
(rank of A) + (nullity of A) = 4,
it follows that the nullity of A is 2.
Column space

Definition. The column space of an m×n matrix

A is the subspace of Rm spanned by columns of A.

Theorem 1 The column space of a matrix A coincides with

the row space of the transpose matrix AT .
Theorem 2 Elementary column operations do not change
the column space of a matrix.
Theorem 3 Elementary row operations do not change the
dimension of the column space of a matrix (although they can
change the column space).
Theorem 4 For any matrix, the row space and the column
space have the same dimension.
Problem. Find a basis for the column space of the
matrix  
1 0 −1 2
B= 2 1 2 3 .
−1 0 1 −2

The column space of B coincides with the row

space of B T . To find a basis, we convert B T to row
echelon form:
   
1 2 −1 1 2 −1
 0 1 0
 → 0 1 0
 

−1 2 1 0 4 0
2 3 −2 2 3 −2
      
1 2 −1 1 2 −1 1 2 −1
0 1 0
 → 0 1 0 → 0 1 0
   
→
0 4 0 0 0 0 0 0 0
0 −1 0 0 −1 0 0 0 0

Thus vectors (1, 2, −1) and (0, 1, 0) form a basis for

the column space of the matrix B.
Problem. Find a basis for the column space of the
matrix  
1 0 −1 2
B= 2 1 2 3 .
−1 0 1 −2

Alternative solution: The dimension of the column

space equals the dimension of the row space, which
is 2 (since the first two rows are not parallel and the
third row is a multiple of the first one).
The 1st and the 2nd columns, (1, 2, −1) and
(0, 1, 0), are linearly independent. It follows that
they form a basis for the column space (actually,
any two columns form such a basis).