Linear Algebra and Differential Equations

Sartaj Ul Hasan

Department of Mathematics
Indian Institute of Technology Jammu
Jammu, India - 181221

Email: sartaj.hasan@iitjammu.ac.in

Sartaj Ul Hasan (IIT Jammu) SMD002U1M 0 / 10

 Lecture 11
(March 31, 2021)

Sartaj Ul Hasan (IIT Jammu) SMD002U1M 1 / 10

Fundamental Results (Conti . . . )
Proposition 13: If V is a finite-dimensional vector space, then any two
bases of V have the same number of elements.
Proof: Suppose B1 and B2 are two distinct bases of V such that |B1 | = m
and |B2 | = n. Then by Proposition 12(a), |B1 | ≤ |B2 | i.e. m ≤ n, since B1
is L. I. and B2 is spanning set. In a similar way, B2 ≤ B1 i. e. n ≤ m.
Hence we get: m = n.
Definition: The dimension of a finite-dimensional space is the
number of elements in a basis for V . This is written dim V .
Remark: Proposition 13 ensures that this is a proper definition.
Special Case: The zero subspace {0} is defined to have dimension 0.
However, it does not have a basis. So our insistence that dim{0} = 0
amounts to saying that the empty set of vectors is a basis of {0}.
Thus the statement that “the dimension of a vector space is the
number of vectors in any basis” holds even for zero space.

Sartaj Ul Hasan (IIT Jammu) SMD002U1M 2 / 10

Examples
dim RnR=n
dim RR = 1 √ and any non-zero real number will act as a basis, e.g.,
{1}, {−8}, { 5}, etc.
dim CC = 1 and any non-zero complex number will act as a basis,
e.g., {1}, {i}, {1 + i}, {5 − 7i}etc.
dim QQ = 1 and any non-zero rational number will act as a basis,
e.g., {1}, {a : a 6= 0 and a ∈ Q}.
dim FF = 1 and any non-zero element in F will act as a basis, e.g.,
{1}, {a : a 6= 0 and a ∈ F}.
dim CR = 2 and the set {1, i} is one of its basis.
dim Rm×n
R = mn, and one basis consists of all m × n matrices with
exactly one entry equal to 1 and all other entries equal to 0. We call
this standard basis for Rm×n
dim Rn [t]R = n + 1, and the set {1, t, t 2 , . . . , t n } is one of its bases
generally known as standard basis for Rn [t].
Sartaj Ul Hasan (IIT Jammu) SMD002U1M 3 / 10
How to find Basis

Proposition 14: Suppose S = {v1 , v2 , . . . , vn } is a linearly independent

set in a vector space V . Suppose v is a vector which is not in Span S.
Then the set obtained by adjoining v to S is linearly independent.

Proof: Suppose v is not in Span S, and consider any expression:

c1 v1 + · · · + cn vn + cv = 0 (1)

If c 6= 0, we can write: cv = −c1 v1 − · · · − cn vn

or c −1 cv = −c −1 c1 v1 − · · · − c −1 cn vn
or v = −c −1 c1 v1 − · · · − c −1 cn vn , contradicting the assumption that v is
not in Span S. Hence c = 0. Then (1) becomes c1 v1 + · · · + cn vn = 0,
and since S is linearly independent, ci = 0 for all i. Thus S ∪ {v } is
linearly independent.

Sartaj Ul Hasan (IIT Jammu) SMD002U1M 4 / 10

How to find Basis (Cont’d)
As a consequence of Proposition 14 and its proof, we get the following:
Proposition 15: Any linearly independent set S in a finite-dimensional
vector space can be expanded to a basis.
Proof: Proof is left as an exercise. The proof proceeds by applying Prop
14 repeatedly; by Prop 12, the process cannot go on indefinitely; it has to
stop, and that stage, a basis has been obtained.
In a similar way, we can get a result which works in the “opposite”
direction:
Proposition 16: Any finite spanning set S in a non-zero vector space can
be contracted to a basis.
Proof: Left as an exercise.
Remark: In view of this proposition, we can say that if a non-zero vector
space V has a finite spanning set S, then it must be finite-dimensional.
Remark: We can regard a basis as either a maximal linearly independent
set, or as a minimal spanning set.
Sartaj Ul Hasan (IIT Jammu) SMD002U1M 5 / 10
Summarizing Results about Dimension

Proposition 17: Let V be a finite-dimensional vector space with

dimension n. Then:
(a) Any subset of V which contains more than n elements is linearly
dependent.
(b) No subset of V which contains less than n vectors can span V .

Remark: Proposition 17 essentially summarizes Proposition 12 and its

consequences in terms of dimension.

Sartaj Ul Hasan (IIT Jammu) SMD002U1M 6 / 10

Dimension of Subspaces

Proposition 18: If W is a proper subspace of a finite-dimensional space

V , then W is also finite-dimensional and 0 < dim W < dim V .
(Note: A proper subspace is a subspace different from the zero subspace
and the entire space. )

Proof: Since W is a proper subspace, it contains a vector w1 6= 0. If w1

spans W , then W is finite-dimensional. If not, there is a vector w2 in W
outside Span w1 , and by adjoining w2 to w1 , we still have a linearly
independent set (by Proposition 14). Continuing in this way, we get a
basis of W with at most dim V elements (upper limit comes because of
Proposition 12). Hence, W is finite-dimensional and dim W ≤ dim V . But
since W is a proper subspace, there is a vector v outside W . Adjoining v
to any basis of W , we still have a linearly independent set. Hence, dim W
must be (strictly) less than dim V , i.e. dim W < dim V .

Sartaj Ul Hasan (IIT Jammu) SMD002U1M 7 / 10

Application of Proposition 18
Question: Is C [a, b], i.e. the space of continuous real-valued functions
defined on an arbitrary closed interval, finite-dimensional or
infinite-dimensional ?
Answer: Infinite-dimensional.
Proof: Suppose BWOC that C [a, b] is finite-dimensional. Now, consider
P[a, b], the set of polynomial functions with domain [a, b]. Clearly,
P[a, b] ⊆ C [a, b] since all polynomial functions are continuous.
Furthermore, it is easy to see that P[a, b] is actually a subspace of C [a, b].
At this stage, we recall the result that R[t] is infinite-dimensional. R[t]
and P[a, b] are not exactly the same space, but they are very similar. So
the proof that R[t] is infinite-dimensional can be applied to P[a, b] with
minor modifications to show that P[a, b] is infinite dimensional. But by
Proposition 18, any subspace of a finite-dimensional vector space must be
finite-dimensional. This is a contradiction. Hence, C [a, b] must be
infinite-dimensional.

Sartaj Ul Hasan (IIT Jammu) SMD002U1M 8 / 10

Additional Notes about P[a, b]
The importance of P[a, b] follows from the following deep result:
Weierstrass Approximation Theorem: Let f ∈ C [a, b] and let any  > 0
be given. Then, there exists a polynomial function p(x) ∈ P[a, b] such
that the distance between p(X ) and f (x) is less than , i.e.
|p(x) − f (x)| <  for all x ∈ [a, b].
Remarks: The above is very useful both in the theory and in practical
applications since polynomials are easy to calculate with. Recall Taylor’s
Theorem:
0 h2 00 hn
f (a + h) = f (a) + hf (a) + f (a) + · · · + f n ()a + Remainder Term.
2! n!
This can be used to determine the approximating polynomial provided two
conditions hold:
(i) f (x) is not just continuous but also has derivatives of all orders, i.e.
f ∈ C ∞ [a, b].
(ii) The Remainder Term converges to 0 as n → ∞ (for small h).
Sartaj Ul Hasan (IIT Jammu) SMD002U1M 9 / 10
Additional Notes about P[a, b] (Cont’d)

Though we can find functions f ∈ C ∞ [a, b], which do not satisfy (ii)
above, many of the functions found in practice, such as exponential,
/ C ∞ [a, b] but
trigonometric, etc do not satisfy (ii). Of course, if f ∈
f ∈ C [a, b], we can use Weierstrass Approximation Theorem directly.

Sartaj Ul Hasan (IIT Jammu) SMD002U1M 10 / 10