T1A/T1B, hmcheungae Math 2121 Tutorial 8 November 1, 2018

If the proof starts from axioms, distinguishes several cases, and takes thirteen lines in
the text book ... it may give the youngsters the impression that mathematics consists
in proving the most obvious things in the least obvious way.————George Polya

Vector Spaces:
A (real) vector space is a nonempty set V with two operations called vector addition and scalar multi-
plication satisfying several conditions. We refer to the elements of V as vectors.
The vector addition operation for V should be a rule that takes two input vectors u, v ∈ V and produces
an output vector u + v ∈ V (i.e., Closure under vector addition) such that

(a) (Commutativity of vector addition:) u + v = v + u.

(b) (Associativity of vector addition:) (u + v) + w = u + (v + w).
(c) (Existence of zero vector:)
There exists a unique zero vector 0 ∈ V with the property that 0 + v = v for all v ∈ V .

The scalar multiplication operation for V should be a rule that takes a scalar input c ∈ R and an input
vector v ∈ V and produces an output vector cv ∈ V (i.e., Closure under scalar multiplication) such that
(a) (Existence of inverse element:) If c = −1 then v + (−1)v = 0.
(b) (Distributivity of scalar multiplication over vector addition:) c(u + v) = cu + cv.

(c) (Distributivity of scalar multiplication over scalar addition:) (c + d)u = cu + du.

(d) (Compatibility of scalar multiplication with field multiplication:) c(dv) = (cd)v for c, d ∈ R.
(e) (Identity of scalar multiplication:) If c = 1 then 1v = v.

• Notation: If V is a vector space and v ∈ V then we define −v = (−1)v and u − v = u + (−v).

• Remark: If V is a vector space then 0v = 0 and c0 = 0 for all c ∈ R and v ∈ V .

Examples of vector spaces include (unless otherwise stated, assume the usual addition and
scalar multiplication):
• Map(X, V ) = {f : X → V }, where V is a vector space.
• Map(X, R): The set of functions f : X → R, where X is a set.

• Rn = Map({1, 2, · · · , n}, R): The n-dimensional Euclidean space.

• R∞ = Map({1, 2, · · · }, R): The set of infinite sequences a = (a1 , a2 , a3 , · · · ) of real numbers ai ∈ R.
• V ∗ : The set of linear map from V to R, which is called the dual space of V .

• R[x]: The set of all polynomials (with real coefficients)

• Pn : The set of all polynomials of degree at most n.
• C[a, b]: The set of all continuous functions defined on [a, b].
• C k [a, b]: The set of all functions continuous on [a, b] with continuous k-th derivative on (a, b).

• Mm×n : The set of all m × n matrices.

T1A/T1B, hmcheungae Math 2121 Tutorial 8 Page 2 of 9

Subspaces:
• Let H be a subset of a vector space V . H is called a subspace of V if it satisfies
1. (Existence of zero vector:) 0 ∈ H.
2. (Closure under vector addition:) If u, v ∈ H, then u + v ∈ H.
3. (Closure under scalar multiplication:) If u ∈ H and c ∈ R, then cu ∈ H.
Note that the other conditions must hold for a subset of V , thus a subspace H is itself another vector space.
• If H ⊆ V is a subspace, then dim H ≤ dim V , and if dim H = dim V , then H = V .
• Suppose U and V are vector spaces. A function f : U → V is linear if

f (u + v) = f (u) + f (v) and f (cv) = cf (v) for all c ∈ R and u, v ∈ U.

If f : U → V is linear, then range(f) and ker(f ) are subspaces, where

range(f ) = {f (x) : x ∈ U } and ker(f ) = {x ∈ U : f (x) = 0}.

• For any vector space V , V and {0} must be subspaces of V , these two are called improper subspaces.
Any other subspace of V is called a proper subspace.

Spans:
Suppose V is a vector space.
• Let v1 , · · · , vk ∈ V , then

Span{v1 , · · · , vk } = {c1 v1 + · · · + ck vk | c1 , · · · , ck ∈ R}.

i.e., The span of v1 , · · · , vk is the set of all linear combinations of v1 , · · · , vk .

– More generally, if S is a (possibly infinite) subset of V , then
( k )
X
Span S = ci vi : k ∈ N, vi ∈ S, ci ∈ R .
i=1

i.e., Span S is the collection of all finite linear combinations of vectors in S.

• Span S is a subspace of V . In particular, if Span S = V , we say that S spans V .

Linear Independence
• Let v1 , · · · , vk ∈ V , then {v1 , · · · , vk } is said to be linearly independent if

x1 v1 + · · · + xk vk = 0 =⇒ x1 = x2 = · · · = xk = 0,

otherwise it is said to be linearly dependent.

– More generally, if S is a (possibly infinite) subset of V , then S is linearly independent if every
finite subset of S is linear independent.

Basis for a Vector Space

Let V be a vector space. We get the following facts:

(a) V has at least one basis.

(b) Every basis of V has the same size.

Cont.
T1A/T1B, hmcheungae Math 2121 Tutorial 8 Page 3 of 9

(c) If A is a subset of linearly independent vectors in V , then V has a basis B with A ⊆ B.

(d) If C is a subset of vectors in V whose span is V , then V has a basis B with B ⊆ C.

We define the dimension of a vector space V to be the common size of any of its bases.
Denote the dimension of V by dim V .

Let U be a subspace of V , B be a collection of vectors in U . Note that B is called a basis for U if

• B spans U ,
• B is linearly independent.

Example 1
True or false:
1. Let Pk be the vector space of all polynomials of degree k or less with real coefficients.
Then S1 = {f (x) ∈ P4 | f (1) is an integer} is a subspace of P4 .
2. S2 = {f (x) ∈ P4 | f (1) is a rational number} is a subspace of P4 .

3. S3 = {A ∈ M2×2 (R) | det(A) 6= 0} is a subspace of M2×2 (R).

4. S4 = {A ∈ M2×2 (R) | det(A) = 0} is a subspace of M2×2 (R)
5. Let C[a, b] be the vector space of all real continuous functions defined on the interval [a, b].
Then S5 = {f (x) ∈ C[−2, 2] | f (−1)f (1) = 0} is a subspace of C[−2, 2].

6. S6 = {f (x) ∈ C[−1, 1] | f (x) ≥ 0 for all − 1 ≤ x ≤ 1} is a subspace of C[−1, 1].

7. Let C 2 [a, b] be the vector space of all real-valued functions f (x) defined on [a, b], where f (x), f 0 (x),
and f 00 (x) are continuous on [a, b]. Here f 0 (x), f 00 (x) are the first and second derivative of f (x).
Then S7 = {f (x) ∈ C 2 [−1, 1] | f 00 (x) + f (x) = sin(x) for all − 1 ≤ x ≤ 1} is a subspace of C[−1, 1].

8. Let S8 be the set of real polynomials of degree exactly k, where k ≥ 1 is an integer, in the vector space
Pk .
Then S8 is a subspace of Pk .
9. If v1 , · · · , vp are in a vector space V , then S9 =Span{v1 , · · · , vp } is a subspace of V .
 
2a b
10. The set S10 of all matrices of the form is a subspace of M2×2 .
3a + b 3b
Answer:
1
1. False: Consider f (x) = x, deg f ≤ 4 and f (1) = 1 ∈ Z, then f ∈ S1 . But f 6∈ S1 .
2
2. False:
√ Let f (x) = x. Then f (x) is a degree 1 polynomial and f (1) = 1 is a rational number. But
2f 6∈ S2 .
3. False: The zero vector of the vector space M2×2 is the 2 × 2 zero matrix O. Since the determinant of
the zero matrix O is 0, it is not in S3 .
   
1 0 0 0
4. False: Consider the matrices A = and B = . The determinants of A and B are both
0 0 0 1
0, hence they belong to S4 . However, their sum A + B = I2 has the determinant 1, hence the sum
A + B 6∈ S4 .

Cont.
T1A/T1B, hmcheungae Math 2121 Tutorial 8 Page 4 of 9

5. False: Consider the continuous functions f (x) = x − 1 and g(x) = x + 1. (These are polynomials,
hence they are continuous.) We have f (−1)f (1) = (−2)(0) = 0 and g(−1)g(1) = (0)2 = 0.
So these functions are in S5 .
However, their sum h(x) := f (x) + g(x) does not belong to S5 since we have h(−1)h(1) = (f (−1) +
g(−1))(f (1) + g(1)) = (−2 + 0)(0 + 2) = −4 6= 0.
6. False:
Let f (x) = x2 , an open-up parabola. Then f (x) is continuous and non-negative for −1 ≤ x ≤ 1. Hence
f (x) = x2 is in S6 .
However, the scalar product (−1)f (x) of f (x) and the scalar −1 is not in S6 since, say, (−1)f (1) =
−1 < 0
7. False: The zero vector of the vector space C 2 [−1, 1] is the zero function θ(x) = 0. The second
derivative of the zero function is still the zero function. Thus θ00 (x) + θ(x) = 0. and since sin(x) is not
the zero function, θ(x) is not in S7 .
8. False: Note that xk + 1, −xk + 1 ∈ S8 but 2 = (xk + 1) + (−xk + 1) 6∈ S8 since deg 2 = 0.
9. True:
• 0 ∈ S9 since 0v1 + · · · + 0vp ∈ S9 .
• Given u, w ∈ S9 . Let u = a1 v1 + · · · + ap vp and w = b1 v1 + · · · + bp vp .
Then u + w = (a1 + b1 )v1 + · · · + (ap + bp )vp ∈ S9 .
• Let c ∈ R, u ∈ S9 . Let u = a1 v1 + · · · + ap vp . Then Let cu = (ca1 )v1 + · · · + (cap )vp ∈ S9
10. True:
         
2a b 2 0 0 1 2 0 0 1
Since =a +b . Then S9 =span , ≤ M2×2 .
3a + b 3b 3 0 1 3 3 0 1 3

Example 2
True or false:
Consider two subsets of C(R), the vector space of continuous functions from R to R.
Let A be a n × n matrix and z ∈ Rn a vector with the property that Ak z 6= 0 but Ak+1 z = 0.

1. S = {2 sin2 x, 3 cos2 x, cos 2x} ⊆ C(R) is linearly independent.

2. S 0 = {x, sin x, cos x} ⊆ C(R) is linearly independent.
3. S 00 = {z, Az, · · · , Ak z} ⊆ Rn are linearly independent.

Answer:

1. False:
From the trigometric identity cos 2x = cos2 x − sin2 x, we get

3(2 sin2 x) − 2(3 cos2 x) + 6(cos 2x) = 0.

Hence, S is linearly dependent.

2. True:
Claim S 0 is linearly independent. To prove this, assume we have scalars c1 , c2 , c3 such that

c1 x + c2 sin x + c3 cos x = 0.

This equation holds for all x ∈ R.

Cont.
T1A/T1B, hmcheungae Math 2121 Tutorial 8 Page 5 of 9

Putting x = 0, we have c3 = 0.
Thus, the equation reduces to

c1 x + c2 sin x = 0.

Putting x = π, we see that c1 = 0.

This leaves us with c2 sin x = 0.
π
Putting x = , we obtain c2 = 0. This implies c1 x = 0, putting x = 1, we get c1 = c2 = c3 = 0.
2
3. True:
Consider

a0 z + a1 Az + · · · + ak Ak z = 0,

then applying Ak on both sides, we get

a0 Ak z + a1 Ak+1 z + · · · + ak A2k z = Ak 0 = 0,

this implies

a0 Ak z = 0,

since Ak z 6= 0, we get a0 = 0.

We get

a1 Az + · · · + ak Ak z = 0,

and applying Ak−1 on both sides, we get a1 = 0. Repeating the process, we get a0 = a1 = · · · = ak = 0.
Therefore, {z, Az, · · · , Ak z} is a linearly independent set.

Example 3
True or False:

1. Let U be a vector space, V and W be subspaces of U . Then V ∩ W is a subspace of U .

2. Let U be a vector space and V , W be its subspaces. Then

V + W = {v + w ∈ U | v ∈ V and w ∈ W }

is a subspace of U .
3. A subset W of a vector space V is a subspace of V if and only if Span(W ) = W .

Answer:

1. True:
We verify by definition.
• 0U ∈ V and 0U ∈ W . Then 0U ∈ V ∩ W .
• Given u1 , u2 ∈ V ∩ W ,

u1 , u2 ∈ V ∩ W ⊆ V =⇒ u1 + u2 ∈ V,

u1 , u2 ∈ V ∩ W ⊆ V =⇒ u1 + u2 ∈ W,
then u1 + u2 ∈ V ∩ W .

Cont.
T1A/T1B, hmcheungae Math 2121 Tutorial 8 Page 6 of 9

• Given u ∈ V ∩ W , and c ∈ R.
Note that u ∈ V ∩ W ⊆ V , then au ∈ V . Similarly, u ∈ V ∩ W ⊆ W , then au ∈ W .
Hence, au ∈ V ∩ W .
Therefore, V ∩ W is a vector subspace of U .
2. True:
• 0U = 0U + 0V ∈ V + W .
• For each x, y ∈ V + W , there are v, v0 ∈ V and w, w0 ∈ W , then

x + y = (v + w) + (v0 + w0 ) = (v + v0 ) + (w + w0 ) ∈ V + W.

• For each x ∈ V + W , and c ∈ R, there exists v ∈ V and w ∈ W such that x = v + w.

Then

cx = cv + cw ∈ V + W.

Therefore,

V + W = {v + w ∈ U | v ∈ V and w ∈ W }

is a subspace of U . This subspace is called the sum of V and W .

3. True:
(⇒) It is clear that W ⊆ Span(W ).
We need to show that if W is a subspace of V , then Span(W ) ⊆ W .
For each u ∈ Span(W ),

u = a1 v1 + a2 v2 + · · · + an vn

for some v1 , v2 , · · · , vn ∈ W and some scalars a1 , a2 , · · · , an ∈ R.

Since W is a subspace of V and v1 , v2 , · · · , vn ∈ W ,

u = a1 v1 + a2 v2 + · · · + an vn ∈ W.

So, Span(W ) ⊆ W .

Hence if W is a subspace of V , then Span(W ) = W

(⇐) By the fact that spanning set is a vector space, we have W = Span(W ) is a subspace of W .

Example 4
Let V and W be vector spaces and T : V → W be a linear transformation. Show that
(a) For each subspace H of V , the direct image T (H) := {T (h) ∈ W : h ∈ H} a subspace of W .
(b) For each subspace K of W , the inverse image T −1 (K) := {x ∈ V : T (x) ∈ K} is a subspace of V .
(c) Im T := {T (v) ∈ W : v ∈ V } ≤ W .
(d) ker T := {v ∈ V : T (v = 0)} ≤ V
Answer:

(a) • Since 0V ∈ H, then 0W = T (0W ) ∈ T (H).

Cont.
T1A/T1B, hmcheungae Math 2121 Tutorial 8 Page 7 of 9

• For each x, y ∈ T (H), there exist a, b ∈ H such that x = T (a) and y = T (b), then

x + y = T (a) + T (b) = T (a + b) ∈ T (H)

since a, b ∈ H implies a + b ∈ H.
• For each x ∈ T (H) and r ∈ R, there is a ∈ H such that x = T (a), then

rx = rT (a) = T (ra) ∈ T (H)

since a ∈ H implies ra ∈ H.
Therefore, T (H) ≤ W .
(b) • Since T (0V ) = 0W , 0V ∈ T −1 (K).
• For each x, y ∈ T −1 (K), then T (y), T (y) ∈ K, then T (x + y) = T (x) + T (y) ∈ K.
• For each x ∈ T −1 (K), and r ∈ R, then T (x) ∈ K and T (rx) = rT (x) ∈ K.
Thus rx ∈ T −1 (K).
Therefore, T −1 (K) ≤ V .
(c) By part (a), taking H = V ≤ V .
(d) By part (b), taking K = {0W } ≤ W .

Example 5
(a) Show that the subspace W consisting of symmetric matrices is a subspace of M2×2 (R).
(Note: W = {A ∈ M2×2 (R) : A = AT }.)
(b) Find a basis for W . What is the dimension of W?
(c) Find a basis for M2×2 (R) which contains a basis for W .

Answer:
 
0 0
(a) Clearly, ∈ W.
0 0
Let w1 , w2 ∈ W .
We need to show that w1 + w2 ∈ W and cw1 ∈ W .
   
x1 y1 x2 y2
Write out w1 = , w2 = .
y1 z1 y2 z2
Then
   
cx1 cy1 x1 + x2 y1 + y2
cw1 = , w1 + w2 =
cy1 cz1 y1 + y2 z1 + z2

is also symmetric, i.e., w1 + w2 ∈ W and cw1 ∈ W .

So W is a subspace of M2×2 (R).
(b) Claim that
     
1 0 0 1 0 0
, ,
0 0 1 0 0 1

is a basis for W

Cont.
T1A/T1B, hmcheungae Math 2121 Tutorial 8 Page 8 of 9

Linear independence: For a1 , a2 , a3 ∈ R, the linear combination

       
1 0 0 1 0 0 0 0
a1 + a2 + a3 =
0 0 1 0 0 1 0 0

Then
   
a1 a2 0 0
=
a2 a3 0 0

Then a1 = a2 = a3 = 0.
 
a1 a2
Spanning: Any symmetric matrix can be represented as a linear combination by
a2 a3
       
a1 a2 1 0 0 1 0 0
= a1 + a2 + a3
a2 a3 0 0 1 0 0 1

So W has a basis of three elements, and has dimension 3.

(c) Extend the basis from (b) to the list

        
1 0 0 1 0 0 0 1
B= , , , .
0 0 1 0 0 1 0 0

Idea:
 
0 1
Note that is asymmetric and
0 0
the span of the first three contains only symmetric matrices.

           
a1 a2 + a4 1 0 0 1 0 0 0 1 0 0
= a1 + a2 + a3 + a4 =
a2 a3 0 0 1 0 0 1 0 0 0 0

This implies a1 = a2 = a3 = a4 = 0.
Hence the four matrices are linearly independent.
Because dim M2×2 (R) = 4, we get B is a basis of M2×2 (R).

More challenging problems

Example i
Let U be a vector space, V and W be subspaces of U . Show that V ∪ W is a subspace of U if and only if
V ⊆ W or W ⊆ V .
Answer:
(⇐) Suppose V ⊆ W or W ⊆ V , then

V ∪ W = V or V ∪ W = W,

which is a vector subspace of U in any case.

(⇒) (Proof by contradiction)
Suppose V ∪ W ≤ U , suppose on contrary that V 6⊆ W or W 6⊆ V .
Then there exists v ∈ V and w ∈ W such that v 6∈ W and w 6∈ V .

Cont.
T1A/T1B, hmcheungae Math 2121 Tutorial 8 Page 9 of 9

Since v, w ∈ V ∪ W and V ∪ W ≤ U , we get

v + w ∈ V ∪ W,

which implies

v + w ∈ V or v + w ∈ W.

If v + w ∈ V , then w = (v + w) − v ∈ V , a contradiction.
If v + w ∈ W , then v = (v + w) − w ∈ W , a contradiction.
Thus V ⊆ W or W ⊆ V .

The End.