MA2102 : Linear Algebra Notes

Group : F

Instructor : Dr. Soumya Bhattacharya

Ratnadeep Sahoo Devansh Joshi Naren Pathania
(23MS232) (23MS166) (23MS248)
Mayank Gohain Shreetam Muduli
(23MS186) (23MS033)

Introduction :
Today we will dive into the notion of Dimension of Finite Dimensional Vector Space,
Relation between the cardinalities of (finite) Spanning sets of V and (finite) linearly
independent subsets of V, Definitions of Basis, Two Propositions on Basis. And at
the end of this lecture, we will define the notion of Coordinates.

Recall :
In the last class, we were introduced with the notions of Basis, Finite and infinite Dimen-
sional Vector Space, some Propositions on Basis and Span.

1
 V : Finite Dimensional Vector Space
|

𝔽 : Field
S := A finite set of vectors which span V
L := A linearly independent set of vectors in V
⇓

|L| ≤ |S|

Now we will continue after that.

1. Collection of linearly independent subsets(L) of V and

Collection of spanning sets(S) of V and their relation :
First we will define two notations.

𝜑 := {S ⊂ V | S spans V}
Γ := {L ⊂ V | L is linearly independent}
Relation :
max{|𝐿|} ≤ min{|𝑆|}
𝐿∈Γ 𝑆∈𝜑

Since |L| ≤ |S|, The cardinality of L will always be less than or equal to the cardi-
nality of S. So, any element which belongs to {L} will be less than or equal to
any element which belongs to {S}. So, it trivially follows that the maximum car-
dinality of set consisting linearly independent subsets(L) of the vector space
V will be less than or equal to the minimum cardinality of set of Spanning sets
of V.

2
Note : Can Strict Inequality be possible in the above relation ?

The strict inequality is not actually possible. We will prove it by the Method of Con-
tradiction.
Let’s assume that 𝐿0 ∈ Γ and,
|𝐿0 | = max{|𝐿|}
𝐿∈Γ

Now if
|𝐿0 | < min{|𝑆|}
𝑆∈𝜑

then, Span (𝐿0 ) ≠ 𝑉

⟹ ∃ v ∈ 𝑉 such that v ∉ Span (𝐿0 )
Let 𝐿′0 ≔ (𝐿0 , 𝑣)
Then, 𝐿′0 is linearly independent
∴ |𝐿′0 | = |𝐿0 | + 1
Which is contradicting with the Maximality of |𝐿0 |
So, there should surely exist an equality instead of strict inequality in this relation.

2. Definition : Dimension of a Finite Dimensional Vector

Space
𝜑 := {S ⊂ V | S spans V}
Γ := {L ⊂ V | L is linearly independent}
If max𝐿∈Γ {|𝐿|} = min𝑆∈𝜑 {|𝑆|} = 𝛼 ∈ ℕ, then the number 𝛼 is called the Dimension
of Finite Dimensional Vector Space, V.
Dimension of Finite Dimensional Vector Space V i.e. dim (V) < ∞

3
3. Definition : Basis of Vector Space

3.1. An ordered linearly independent set of vectors in V of cardi-

nality equal to dim (V) is known as a Basis of Vector Space V.
3.2. An ordered set of vectors in Vector Space V which spans V
and has cardinality equal to dim (V) is familiar as a Basis of Vec-
tor Space V.

4. Some Propositions :

Proposition 4.1 : If 𝔹 and 𝔹′ are two basis of a finite dimensional vector space
V, then
|𝔹| = |𝔹′ |

4
Proof :
We know that if
S := A finite set of vectors which span V
L := A linearly independent set of vectors in V
Then, |L| ≤ |S|
Case 1 : Let’s take L = 𝔹 and S = 𝔹′
∴ |𝔹| ≤ |𝔹′ | − −−−−−−−−−(1)

Case 2 : Let’s take L = 𝔹′ and S = 𝔹

∴ |𝔹′ | ≤ |𝔹| − −−−−−−−−−(2)

From (1) and (2), we get

|𝔹| = |𝔹′ |

Hence Proved Proposition (4.1)

◾

Proposition 4.2 : Let V be a finite dimensional vector space and let W be a

subspace of V. Then dim (W) ≤ dim(V), where the equality holds if W = V.

Proof : First we will show that Every subspace of a finite dimensional vector space is
finite dimensional.
Let’s Suppose V is finite dimensional vector space and U is a subspace of V, We will
prove U is finite dimensional through the following multistep construction.
• Step 1 : If U = {0}, then U is finite dimensional and we are done. If U ≠ {0}, then
we have to choose a nonzero vector 𝑣1 ∈ U.
• Step j : If U = span(𝑣1 , 𝑣2 , …, 𝑣𝑗−1 ), then U is finite dimensional and we are done. If U ≠
span(𝑣1 , 𝑣2 , …., 𝑣𝑗−1 ), then we have to choose a vector 𝑣𝑗 such that 𝑣𝑗 ∉ span(𝑣1 , 𝑣2 , …, 𝑣𝑗−1 )
After each step, as long as the process continues, we have constructed a set of vectors such
that no vector in this set is in the span of the previous vectors. Thus after each step,
we have constructed a linearly independent set. This linearly independent set cannot be
longer than any spanning set of V, and thus the process must eventually terminate, which
means that U is finite dimensional.
So, W is finite dimensional and hence has a basis as W is a subspace of finite dimen-
sional vector space V.
Let 𝔹 be a basis of W. If Span (𝔹) = V ⟺ W = V.
Otherwise, ∃ v ∈ V such that v ∉ W
∴ v ∉ Span (𝔹)
⇒ 𝔹′ ≔ (𝔹,v)

5
 ∴ 𝔹′ is linearly independent.
Since, 𝔹′ is a linearly independent subset of V, |𝔹′ | ≤ dim (V).
Therefore, the process of adding elements to 𝔹 must come to an end. Since 𝔹 contains at
most dim (V) elements,
∴ dim (W) ≤ dim (V)
Now, we will show that if dim(W)=dim(V) then W = V. B has the cardinality equal to
dim (W) = dim (V). Since B is linearly independent and B is a basis of V ⇒ W =
span(B) = V.
Hence Proved Proposition (4.2)
◾

5. Definition : Coordinates
Let 𝔹 = (𝑣1 , 𝑣2 , …., 𝑣𝑛 ) be a basis of Vector Space V. If each 𝑣 ∈ 𝑉 can be ex-
𝑐1
⎛ 𝑐2 ⎞
pressed as 𝑣 = 𝑐1 𝑣1 + 𝑐2 𝑣2 + …… + 𝑐𝑛 𝑣𝑛 , then ⎜
⎜⋮⎟ ⎟ is called the coordinates of 𝑣
𝑐
⎝ 𝑛⎠
with respect to 𝔹.

Note : When The basis change occurs, the coordinates also change simultaneously.

Remark: Does the basis of a non-empty finite-dimensional vector space need to be

ordered?

The usual definition of bases in most linear algebra-related books do not include any ordering.
The best way to understand why is by explicitly introducing ordered bases.

6
 Let 𝑆 be a finite non-empty set of vectors in 𝑉 . Then-
• 𝑆 is a spanning set of 𝑉 consisting of linearly independent vectors in 𝑉 .
• Every non-zero vector 𝑣 ∈ 𝑉 is a unique linear combination of vectors in 𝑆.
A set of vectors in 𝑉 that satisfy any (and hence both) of these conditions is called a basis
for 𝑉 .

In neither of these conditions is the notion of order explicitly mentioned to append the defini-
tion of basis. Further, to introduce the concept of coordinate matrices, it will be convenient to
consider bases that have an order imposed on their members.

Let 𝑉 be a vector space of dimension 𝑛. An ordered basis for 𝑉 is an ordered 𝑛-

tuple (𝑣1 , 𝑣2 , …, 𝑣𝑛 ) of vectors for which the set {𝑣1 , 𝑣2 , …, 𝑣𝑛 } is a basis for 𝑉 . If 𝐵 =
(𝑣1 , 𝑣2 , …, 𝑣𝑛 ) is an ordered basis for 𝑉 , then for each 𝑣 ∈ 𝑉 there exists a unique ordered
𝑛-tuple (𝑟1 , 𝑟2 , …𝑟𝑛 ) of scalars for which-
𝑣 = 𝑟1 𝑣1 + 𝑟2 𝑣2 + … + 𝑟𝑛 𝑣𝑛
Accordingly we can define the coordinate map 𝜙𝐵 : 𝑉 → 𝔽𝑛 by
𝑟1
⎛ 𝑟2 ⎞
𝜙𝐵 (𝑣) = (𝑣)𝐵 = ⎜
⎜⋮⎟ ⎟
𝑟
⎝ ⎠ 𝑛

∴ There is a clear difference between ordered basis and basis. The suggestive index notation of
{𝑣1 , 𝑣2 , …, 𝑣𝑛 }
does not imply order as we have to additionally append the concept of an 𝑛-tuple (𝑣1 , 𝑣2 , …, 𝑣𝑛 ).
What’s crucial however, is how we define coordinates via 𝑛-tuples, which are ordered objects. So
whenever we do something related to coordinate maps we need to explicitly or implicitly define
basis to respect the 𝑛-tuple of coordinates. The suggestive index notation is merely convenient
to add an ordered description for the basis if necessary (this is used for a non-explicit introduc-
tion to ordered bases. For further reading on the same refer to Serge Lang’s introduction to
coordinates)
Basically, a basis is defined without the notion of ordering; but whenever coordinates come into
play, they impose an ordering on the basis.

7
6. Questions :

1. [Ratnadeep] : Let S = {(5,1,1), (1,5,1), (1,1,5)}. Can S be a basis of ℝ3 ? Give justi-

fication.
2. [Ratnadeep] : Let 𝑢 and 𝑣 be distinct vectors of a vector space 𝑉 . If {𝑢, 𝑣} is a basis
for 𝑉 , then show that both {𝑢 + 𝑣, 𝑎𝑢} and {𝑎𝑢, 𝑏𝑣} are also bases for V, where 𝑎 and 𝑏
are non-zero scalars.
3. [Ratnadeep] : Let V be a real vector space with {𝛼, 𝛽, 𝛾} as a basis. Prove that the set
{𝛼 + 𝛽 + 𝛾, 𝛽 + 𝛾, 𝛾} is also a basis of V.
4. [Ratnadeep] : Find a basis for the vector space ℝ3 containing the vectors (1, 5, 1)
and (3, 5, 6).
5. [Ratnadeep] : If {𝛼1 , 𝛼2 , …., 𝛼𝑛 } be a basis of Vector space V over Field 𝔽 and a
non-null vector 𝛽 in V can be expressed as 𝛽 = 𝑐1 𝛼1 + 𝑐2 𝛼2 + …… + 𝑐𝑛 𝛼𝑛 , 𝑐𝑖 ∈ 𝔽, then
{𝛼1 , 𝛼2 , …., 𝛼𝑗−1 , 𝛽, 𝛼𝑗+1 , ….., 𝛼𝑛 } will be a new basis of V if 𝑐𝑗 ≠ 0.
6. [Ratnadeep] : Let {𝑥1 , 𝑥2 , …., 𝑥𝑛 } be a basis of vector space V over the field 𝔽 and
also let 𝑎𝑖𝑗 ∈ 𝔽, 1 ≤ 𝑖 ≤ 𝑗 ≤ 𝑛. Show that
𝑦1 = 𝑎11 𝑥1 , 𝑦2 = 𝑎12 𝑥1 + 𝑎22 𝑥2 , ….., 𝑦𝑛 = 𝑎1𝑛 𝑥1 + 𝑎2𝑛 𝑥2 + …. + 𝑎𝑛𝑛 𝑥𝑛

is a basis of V iff 𝑎11 𝑎22 …𝑎𝑛𝑛 ≠ 0.

7. [Ratnadeep] : Let V be a finite dimensional vector space of dimension n over a
field 𝔽. Show that No subset of V which contains less than n vectors can span V.
8. [Ratnadeep] : Let 𝑊1 and 𝑊2 be subspaces of a vector space V having dimensions
𝑚 and 𝑛 respectively, where 𝑚 ≥ 𝑛. Prove that -
• dim (𝑊1 ∩ 𝑊2 ) ≤ 𝑛.
• dim (𝑊1 + 𝑊2 ) ≤ 𝑚 + 𝑛
9. [Ratnadeep] : Let 𝑊1 , 𝑊2 be two subspaces of a finite dimensional vector space 𝑉 .
Prove that dim (𝑊1 ) + dim(𝑊2 ) = dim (𝑊1 ∩ 𝑊2 ) + dim (𝑊1 + 𝑊2 ).
10. [Ratnadeep] : Find the coordinate vector of 𝛼 = (1, 3, 1) relative to the ordered basis
𝔹 = (𝛼1 , 𝛼2 , 𝛼3 ) of ℝ3 , where 𝛼1 = (5, 4, 8), 𝛼2 = (3, 2, 0), 𝛼3 = (5, 0, 0).
11. [Devansh] : Assume that any 3 of the five vectors 𝑣𝑗 = (𝑥𝑗 , 𝑦𝑗 , 𝑧𝑗 ) in ℝ3 , j = 1,…,5
are linearly independent. Prove that the vectors

𝑤𝑗 = (𝑥2𝑗 , 𝑦𝑗2 , 𝑧𝑗2 , 𝑥𝑗 𝑦𝑗 , 𝑥𝑗 𝑧𝑗 , 𝑦𝑗 𝑧𝑗 )

are linearly independent in ℝ6 .

12. [Naren] : For a vector space formed by polynomial of degree at most 5, how many
bases can be formed for this vector space?
13. [Naren] : For a vector space formed by polynomial of degree at most 5, give the
examples of at most all or at least 3 bases for this vector space if multiple bases are
possible.
14. [Mayank] : Consider the Vector space ℂ3 over ℂ, find the coordinate matrix of 𝑣 =
(1, 0, 1) in the basis of ℂ3 consisting of the vectors (𝑖, 0, 1), (2, −1, 𝑖), (1, 1 + 𝑖, 0).

8
15. [Mayank] : Consider V as a vector space of real numbers over the field of rational
numbers, with the usual operations. Prove that this vector space is not finite di-
mensional.
16. [Shreetam] : Let 𝑊 ⊂ ℝ4 be the solutions of the system of linear equation 𝐴𝑋 =
0, where
2 1 2 3
𝐴 = (1 1 3 0). Find a basis for 𝑊 .