Math 115a: Selected Solutions for HW 2

October 15, 2005

Exercise 1.4.10: Show that if

     
1 0 0 0 0 1
M1 = , M2 = , M3 = ,
0 0 0 1 1 0

then the span of {M1 , M2 , M3 } is the set of all symmetric 2 matrices.

Solution: Let M be an arbitrary symmetric 2 × 2 matrix; we will denote

 
a b
M= .
b d

Via a rather superficial inspection, we see that

M = aM1 + dM2 + bM3 .

Since we’ve written an arbitrary symmetric matrix as a linear combination of

the Mi0 s, we conclude that {M1 , M2 .M3 } spans our space in question.

Exercise 1.4.14: Show that if S1 and S2 are arbitrary subsets of a vector

space V , then span(S1 ∪ S2 ) = span(S1 ) + span(S2 ). (The sum of two subsets
is defined in the exercises in Section 1.3).

Solution: In order to prove equality of two sets, we need to prove mutual inclu-
sion.
⊆: Let v ∈ span(S1 ∪ S2 ). Then v can be written as a linear combination of
vectors in S1 ∪ S2 , i.e. X X
v= ai xi + bj yj ,
i j

where ai , bj ∈ F and xi ∈ S1 , yj ∈ S2 . (We note that P the two sums are finite,
although
P we will not use this fact in this proof.) Since i ai xi ∈ span(S1 ), and
j bj yj ∈ span(S2 ), we conclude that v ∈ span(S1 ) + span(S2 ).

⊇: Let v ∈ span(S1 ) + span(S2 ). Then by definition,

X X
v= ai xi + bj yj ,
i j

1
where ai , bj ∈ F and xi ∈ S1 , yj ∈ S2 . This is clearly a linear combination of
vectors from S1 ∪ S2 . Therefore v ∈ span(S1 ∪ S2 ).

Exercise 1.5.15: Let S = {u1 , u2 , ..., un } be a finite set of vectors. Prove that
S is linearly dependent if and only if u1 = 0 or uk+1 ∈ span({u1 , u2 , ..., uk }) for
some k (1 ≤ k < n)

Proof:
(⇒) Suppose that S is linearly dependent. Then we need to prove that ei-
ther u1 = 0 or there exists some k such that uk+1 ∈ span({u1 , u2 , ..., uk }). If
u1 = 0, then we are done. So let us suppose that u1 6= 0. Then what we need
to prove is that the second part of the statement must be true: there exists
some k such that uk+1 ∈ span({u1 , u2 , ..., uk }). The way we approach this is to
proceed via proof by contradiction. Suppose that there is no such k such that
uk+1 ∈ span({u1 , u2 , ..., uk }). To rephrase, this means that for all k, uk+1 ∈ /
span({u1 , u2 , ..., uk }). So we now need to use this assumption repeatedly, as
follows: u2 ∈ / span({u1 }) implies that {u1 , u2 } is a linearly independent set.
Similarly, u3 ∈ / span({u1 , u2 }) implies that {u1 , u2 , u3 } is a linearly independent
set. We can continue in this fashion until we get the following final statement:
un ∈/ span({u1 , u2 , ..., un−1 }) implies that S = {u1 , u2 , ..., un } is a linearly de-
pendent set. However, our initial assumption is that S is linearly dependent.
This is our contradiction. Therefore our initial assumption is false; there must
exist some k such that uk+1 ∈ span({u1 , u2 , ..., uk }).

(⇐) Suppose that u1 = 0 or uk+1 ∈ span({u1 , u2 , ..., uk }) for some k (1 ≤

k < n). If u1 = 0, then that means 0 ∈ S, which immediately implies that S
is linearly dependent (why?) So suppose that u1 6= 0. This means that there
exists some k such that u1 = 0 or uk+1 ∈ span({u1 , u2 , ..., uk }). Therefore
T = {u1 , u2 , ..., uk+1 } is a linearly dependent set. Since T ⊆ S, this implies
that S is linearly dependent.

Exercise 1.6.12: Let u, v, w be distinct vectors of a vector space V . Show

that if {u, v, w} is a basis for V , then {u + v + w, v + w, w} is also a basis for V .

Solution: Let {u, v, w} be a basis for V . Since this is a three element set, we
conclude that the dimension of V must be 3. Looking at {u + v + w, v + w, w},
we see that this is also a three element set. Therefore if we can prove that this
set is either linearly independent or spans V , then we are done (make sure you
understand why this is true). We will show that {u + v + w, v + w, w} is a
linearly independent set. Let a1 , a2 , a3 ∈ F such that

a1 (u + v + w) + a2 (v + w) + a3 (w) = 0.

We will show that this implies that a1 = a2 = a3 = 0, by rewriting the equality

2
as:
0 = a1 (u + v + w) + a2 (v + w) + a3 (w)
= (a1 )(u) + (a1 + a2 )(v) + (a1 + a2 + a3 )(w).
Since {u, v, w} is a basis for V , it is a linearly independent set. Therefore from
the last equality, we can conclude that a1 = a1 + a2 = a1 + a2 + a3 = 0, and
from here we can conclude immediately that a1 = a2 = a3 = 0. Therefore we’ve
proven that {u + v + w, v + w, w} is a linearly independent set. Therefore it is
a basis for V .

Exercise 1.6.20: Let V be a vector space having dimension n, and let S

be a subset of V that generate V .
(a) Prove that there is a subset of S that is a basis for V .
(b) Prove that S contains at least n elements.

Solution:
(a): Let {β1 , β2 , ..., βn } be a basis for V . Since span(S)=V , each of the βi ’s can
be written as a finite linear combination of elements from S. More specifically,
X
β1 = a1,i si
i∈I1
X
β2 = a2,i si
i∈I2
..
.
X
βn = an,i si
i∈In

where all of the aj,i ’s are scalars, and In ’s are finite index sets (see the note at
the end of the proof). Let us define the set
n
[
J= Ij
j=1

be the finite union of all the index sets. Now consider the subset of the vector
space [
T = sj .
j∈J

Since T is a set made up of elements from S, T ⊆ S. Since J is a finite index

set, T is also a finite set. Furthermore, we have constructed this set T that
contains elements from S which ”builds” each of the βi ’s. Therefore
{β1 , β2 , ..., βn } ⊆ span (T )
⊆ span (S) = V
⇒
V = span ({β1 , β2 , ..., βn }) ⊆ span (span (T )) = span (T ) ⊆ V
⇒ span(T ) = V.

3
Since T spans V, and it is a finite set, by the Replacement Theorem (1.10) we
can find a subset of T –call it B–that is a basis for V . It is clear that B is a
subset of S, as it is a subset of T . This finishes our proof.