Homework 1 Solutions

Josh Hernandez
October 25, 2009

1.1 - Introduction
2. Find the equations of the lines through the following pairs of points in space.
b. (3, -2, 4) and (-5, 7, 1)
Solution:
x = (3, -2, 4) + r[(-5, 7, 1) − (3, -2, 4)] = (3, -2, 4) + r(-8, 9, -3).
d. (-2, -1, 5) and (3, 9, 7)
Solution:
x = (-2, -1, 5) + r[(3, 9, 7) − (-2, -1, 5)] = (-2, -1, 5) + r(5, 10, 2).

3. Find the equations of the planes containing the following points in space.
b. (3, -6, 7), (-2, 0, -4), and (5, -9, -2).
Solution:

x = (3, -6, 7) + r[(-2, 0, -4) − (3, -6, 7)] + s[(5, -9, -2) − (3, -6, 7)]
= (3, -6, 7) + r(-5, 6, -11) + s(2, 3, -9)
d. (1, 1, 1), (5, 5, 5), and (-6, 4, 2).
Solution:

x = (1, 1, 1) + r[(5, 5, 5) − (1, 1, 1)] + s[(-6, 4, 2) − (1, 1, 1)]

= (1, 1, 1) + r(4, 4, 4) + s(-7, 3, 1)

7. Prove that the diagonals of a parallelogram bisect each other.

Solution: A parallelogram ABCD in R2 may be translated so that vertex A lies on the origin.
After translation, denote the coordinate-vectors of the adjacent vertices B and D by v and w.
The four vertices therefore have coordinate-vectors 0, v, v + w, and w.
By the midpoint rule, segment BD has midpoint at 12 (v+w), and AC has midpoint 12 (0+(v+w)).
These midpoints are identical, so BD and AC bisect one another.

1.2 - Vector Spaces

2. Write the zero vector of M3×4 (F).

1
 0 0 0 0
Solution: Consider the matrix 0 := 0000 . For any matrix A ∈ M3×4 ,
0000
0 0 0 0  a11 a12 a13 a14
  0+a110+a12 0+a13 0+a14   a11 a12 a13 a14

0+A= 0000 + a21 a22 a23 a24
a31 a32 a33 a34
= 0+a21 0+a22 0+a23 0+a24 = a21 a22 a23 a24
a31 a32 a33 a34
=A
0000 0+a31 0+a32 0+a33 0+a34

Thus 0 is the zero vector of M3×4 (F).

12. Prove that the set of even functions from R to R is a vector space over F (under the usual addition
and scalar multiplication).
Solution: Note that (1) f = g means f (x) = g(x) for all x ∈ R, and (2) the “usual addition and
scalar multiplication” means that (f + g)(x) = f (x) + g(x) and (cf )(x) = cf (x) for all x ∈ R.
The even functions are those f : R → R which satisfy f (-x) = f (x) for all x ∈ R. Adding two
such functions:
(f + g)(-x) = f (-x) + g(-x) = f (x) + g(x) = (f + g)(x),
so the sum f + g is an even function. As for scalar multiplication,

(cf )(-x) = cf (-x) = cf (x) = (cf )(x).

The remaining properties hold equally well for any set S, field F, scalars a, b ∈ F, point x ∈ S and
functions f, g, h ∈ F(S, F).
VS1 Commutativity of addition is inherited from F:

(f + g)(x) = f (x) + g(x) = g(x) + f (x) = (g + f )(x).

VS2 Associativity of addition is inherited from F:

[(f + g) + h](x) = (f + g)(x) + h(x)

= (f (x) + g(x)) + h(x) = f (x) + (g(x) + h(x))
= f (x) + (g + h)(x) = [f + (g + h)](x).

VS3 Define 0 by the mapping 0(x) = 0. Then

(f + 0)(x) = f (x) + 0(x) = f (x).

VS4 Define -f by the mapping (-f )(x) = -(f (x))P. Then

(f + -f )(x) = f (x) + (-f )(x) = f (x) + -f (x) = 0 = 0(x).

VS5 Briefly, (1f )(x) = 1f (x) = f (x).

VS6 Associativity of scalars is inherited from F:

((ab)f )(x) = (ab)f (x) = a(bf (x)) = a((bf )(x)) = (a(bf ))(x).

VS7 Distributivity(1) is inherited from F:

(a(f + g))(x) = a(f + g)(x)

= a(f (x) + g(x)) = af (x) + ag(x)
= (af )(x) + (ag)(x) = (af + ag)(x).

2
 VS8 Distributivity(2) is inherited from F:

((a + b)f )(x) = (a + b)f (x) = af (x) + bf (x)

= (af )(x) + (bf )(x) = (af + bf )(x).

16. Let V = Mm×n (R). Let F = Q, the field of rational numbers. Is V a vector space over F (under the
usual addition and scalar multiplication)?
Solution: Adding two matrices in V, (A + B)ij = Aij + Bij ∈ R, so the matrix sum is in V.
Likewise, (cA)ij = cAij ∈ R, for any c ∈ Q ⊆ R.
VS1 Commutativity of addition is inherited from R:

(A + B)ij = Aij + Bij = Bij + Aij = (B + A)ij

VS2 Associativity of addition is inherited from R:

((A + B) + C)ij = (A + B)ij + Cij

= (Aij + Bij ) + Cij = Aij + (Bij + Cij )
= Aij + (B + C)ij = (A + (B + C))ij

VS3 Define 0 by 0ij = 0 for all i, j. Then

(A + 0)ij = Aij + 0ij = Aij + 0 = Aij

VS4 Define -A by (-A)ij = -Aij . Then

(A + -A)ij = Aij + (-A)ij = Aij + -Aij = 0 = 0ij .

VS5 Briefly, (1A)ij = 1Aij = Aij .

VS6 Associativity of multiplication is inherited from R:

((ab)A)ij = (ab)Aij = a(bAij ) = a((bA)ij ) = (a(bA))ij .

VS7 Distributivity(1) is inherited from R:

(a(A + B))ij = a(A + B)ij = a(Aij + Bij ) = aAij + aBij = (aA)ij + (aB)ij = (aA + aB)ij .

VS8 Distributivity(2) is inherited from R:

((a + b)A)ij = (a + b)Aij = aAij + bAij = (aA)ij + (bA)ij = (aA + bA)ij .

18. Let V = {(a1 , a2 ) : a1 , a2 ∈ R}. For (a1 , a2 ), (b1 , b2 ) ∈ V and c ∈ R, define

(a1 , a2 ) + (b1 , b2 ) = (a1 + 2b1 , a2 + 3b2 ) and c(a1 , a2 ) = (ca1 , ca2 )

Is V a vector space over R with these operations?

3
 Solution: No. Consider the points x = (0, 0) and y = (1, 1). Then

x + y = (0, 0) + (1, 1) = (0 + 2(1), 0 + 3(1)) = (2, 3)

whereas
y + x = (1, 1) + (0, 0) = (1 + 2(0), 1 + 3(0)) = (1, 1).
This additive relation is not commutative, so V is not a vector space.

19. Let V = {(a1 , a2 ) : a1 , a2 ∈ R}. Is V a vector space under the usual addition and with scalar
multiplication defined as (
(0, 0) if c = 0
c(a1 , a2 ) = a2
(ca1 , c ) if c 6= 0

Solution: No. Let a = b = 1, and let x = (1, 1). Then

(a + b)x = 2x = (2(1), 12 ) = (2, 12 ),

whereas
ax + bx = 1x + 1x = (1(1), 11 ) + (1(1), 11 ) = (2, 2).
These relations do not satisfy distributivity(1).

20. Let V be the set of sequences {an } of real numbers. For {an }, {bn } ∈ V and t ∈ R, define
{an } + {bn } = {an + bn } and t{an } = {tan }
Prove that, under these operations, V is a vector space over R.
Solution: For {an }, {bn } ∈ V and s, t ∈ R, {an } + {bn } = {an + bn }, and c{an } = {can }, both of
which are sequences of real numbers.
VS1 Commutativity of addition is inherited from R:

{an } + {bn } = {an + bn } = {bn + an } = {an } + {bn }

VS2 Associativity of addition is inherited from R:

({an } + {bn }) + {cn } = {an + bn } + {cn }

= {(an + bn ) + cn } = {an + (bn + cn )}
= {an } + {bn + cn } = {an } + ({bn } + {cn }).

VS3 Define 0 := {0n } by 0n = 0 for all n. Then

{an } + 0 = {an + 0n } = {an + 0} = {an }

VS4 Define -{an } := {zn } by zn = -an for all n. Then

{an } + -{an } = {an + zn } = {an + -an } = {0} = 0.

VS5 Briefly, 1{an } = {1an } = {an }.

VS6 Associativity of multiplication is inherited from R:

(cd){an } = {(cd)an } = {c(dan )} = c{dan } = c(d{an }).

4
 VS7 Distributivity(1) is inherited from R:

c({an } + {bn }) = c{an + bn } = {c(an + bn )} = {can + cbn } = {can } + {cbn } = c{an } + c{bn }.

VS8 Distributivity(2) is inherited from R:

(c + d){an } = {(c + d)an } = {can + dan } = {can } + {dan } = c{an } + d{an }.

We may also prove this using the method in the note below. Given f ∈ F(N, R), a real-valued
function on the natural numbers, define T (f ) = {f (n)}. T is linear:

T (f + cg) = {(f + cg)(n)} = {f (n) + cg(n)} = {f (n)} + {cg(n)} = {f (n)} + c{g(n)} = T (f ) + cT (g).

Given a sequence {an }, define f = T -1 ({an }) by the mapping f (n) = an . Thus

T (T -1 ({an })) = T (f ) = {f (n)} = {an },

so T is onto. Thus the set of real sequences is a vector space.

Note on section 1.1

You might have noticed that the proofs for problems 12, 16, and 20 were basically the same. We can save
a lot of time by relating these spaces to a known vector space, F(I, F), the space of F-valued functions on
some indexing set I.
For instance, we can think of matrices as functions on the indexing set Im×n = {(1, 1), (1, 2), . . . , (m, n)}.
Each set of indices has a corresponding F-value, so we can treat the indexing set like the domain of an
F-valued function. Now, define T : F(Im×n , F) → Mm×n by the mapping

(T (f ))ij = f ((i, j))

That is, the ijth element of the matrix T (f ) is equal to the value of the function f at the point (i, j). This
is linear, since

(T (f + cg))ij = (f + cg)((i, j)) = f ((i, j)) + cg((i, j)) = (T (f ))ij + c(T (g))ij .

This is onto, since we can easily reverse the mapping, sending a given matrix to a corresponding function.

(T -1 M )((i, j)) = Mij .

The point of this is, now Mm×n is the range of a linear function applied to a vector space. It therefore must
also be a vector space. Instead of verifying all those axioms, we defined a single mapping.
This very useful fact has a rather onerous proof, since section 1.1 is so sparse on theorems. However, it
only needs to be proved once, and an be used with all manner of spaces.

5
 Theorem: Suppose V is a vector space over F, S is a set with a binary addition and
F−multiplication relations (not necessarily closed), and T : V → S is an surjection (onto
function) such that T (u + cv) = T (u) + cT (v) for all u, v ∈ V and c ∈ F. Then S is a vector space
under the given addition and multiplication.

Proof: Given x, y ∈ S, there are u, v in V such that T (u) = x and T (v) = y (since T is
onto). By closure of V,
x + y = T (u) + T (v) = T (u + v) ∈ S
and for any c ∈ F,
cx = cT (u) = T (cu) ∈ S.
VS1 Commutativity of addition is inherited from V:

x + y = T (u) + T (v) = T (u + v) = T (v + u) = T (v) + T (u) = y + x.

VS2 Associativity of addition is inherited from V:

(x + y) + z = (T (u) + T (v)) + T (w) = T (u + v) + T (w)

= T ((u + v) + w) = T (u + (v + w))
= T (u) + T (v + w) = T (u) + (T (v) + T (w)) = x + (y + z).

VS3 Define 0 = T (0). Then

x + 0 = T (u) + T (0) = T (u + 0) = T (u) = x.

VS4 Define -x = T (-u). Then

x + -x = T (u) + T (-u) = T (u + -u) = T (0) = 0.

VS5 Briefly, 1x = 1T (u) = T (1u) = T (u) = x.

VS6 Associativity of scalar multiplication is inherited from V:

(ab)x = (ab)T (u) = T ((ab)u) = T (a(bu)) = aT (bu) = a(bT (u)) = a(bx).

VS7 Distributivity(1) is inherited from V:

a(x+y) = a(T (u)+T (v)) = aT (u+v) = T (a(u + v)) = T (au + av) = aT (u)+aT (v) = ax+ay.

VS8 Distributivity(2) is inherited from V:

(a + b)x = (a + b)T (u) = T ((a + b)u) = T (au + bu) = aT (u) + bT (u) = ax + bx.

This is related to the proof that “the range of a linear transform is a subspace of the codomain”, but that
proof assumes that the codomain of the linear function is already a vector space.

1.3 - Subspaces
2. Determine the transpose (and trace, if applicable) of each of the matrices that follow.

6
    
10 0 -8 10 2 -5
d. A =  2 -4 3  Solution: At =  0 -4 7  , tr(A) = 10 + -4 + 6 = 12.
-5 7 6 -8 3 6
 
  -2 7
-2 5 1 4 5 0
f. Solution: At = 
 1 1 . No trace - matrix is not square.

7 0 1 -6
4 -6

5. Prove that A + At is symmetric for any square matrix A.

Solution: Observe that

(A + At )t = At + (At )t = At + A = A + At .

The matrix A + At is equal to its transpose, and is therefore symmetric.

8. Determine whether the following sets are subspaces of R3 under the operations of addition and scalar
multiplication defined on R3
b. W2 = {(a1 , a2 , a3 ) ∈ R3 : a1 + a3 = 2}
Solution: No. Observe that v = (1, 0, 1) lies in W2 , but 2v = (2, 0, 2) does not.
d. W4 = {(a1 , a2 , a3 ) ∈ R3 : a1 − 4a2 − a3 = 0}
Solution: Yes. Given u = (u1 , u2 , u3 ) and (v1 , v2 , v3 ) in W4 , and r ∈ R, consider u + rv:

(u1 + rv1 ) − 4(u2 + rv2 ) − (u3 + rv3 ) = (u1 − 4u2 − u3 ) + (rv1 − 4rv2 − rv3 )
= (u1 − 4u2 − u3 ) + r(v1 − 4v2 − v3 ) = 0 + r0 = 0

Thus u + rv lies in W4 .
f. W6 = {(a1 , a2 , a3 ) ∈ R3 : 5a21 − 3a22 + 6a23 = 0}
√ √
Solution: No. Let v = (0, 2, 1) and let w = (0, 2, -1).
√ These both lie in W6 , but
w + v = (0, 2 2, 0) does not.

11. Is the set W = {f (x) ∈ P(F ) : f (x) = 0 or f (x) has degree n} a subspace of P(F ) if n ≥ 1?
Solution: No. Consider the polynomials p = xn + xn−1 and q = -xn , both of degree n, so they
lie in W. Their sum is xn−1 , which does not lie in W.

13. Let S be a nonempty set and F a field. Prove that for any s0 ∈ S, Vs0 = {f ∈ F(S, F ) : f (s0 ) = 0} is
a subspace of F(S, F ).
Solution: Fix s0 ∈ S. Given f, g ∈ Vs0 and c ∈ F, observe that

(f + cg)(s0 ) = f (s0 ) + (cg)(s0 ) = f (s0 ) + cg(s0 ) = 0.

Thus f + cg lies in Vs0 .

15. Is the set of all differentiable real-valued functions on R a subspace of C(R)?

Solution: Every differentiable function must be continuous, so the differentiable functions com-
pose a subset of the C(R). The linearity property of derivatives (you can take this for granted,
but it must be mentioned) states that, given f, g differentiable real-valued functions on R, the
derivative (f + cg)0 exists and is equal to f 0 + cg 0 . Thus f + cg is differentiable, and so the
differentiable functions compose a subspace of C(R).

7
20. Prove that if W is a subspace of a vector space V and w1 , w2 , . . . , wn are in W, then a1 w1 + a2 w2 +
· · · + an wn ∈ W for any scalars a1 , . . . , an .
Solution: Of course this is true for a single vector. We will prove, by induction, that it is true for
any finite collection:
Suppose this fact is known for collections of n − 1 vectors in W. Then

a1 w1 + a2 w2 + · · · + an wn = (a1 w1 + · · · + an−1 wn−1 ) + an wn = w + v.

This is a sum of two vectors in W, so it must lie in W as well. Thus the fact is known for collections
of n vectors in W.
21. Show that the set of convergent sequences {an } is a subspace of the vector space of sequences of real
numbers.
Solution: The linearity property of limits (again, you can quote this without proof) states that
if the sequence {an } converges to a, and {bn } converges b, then {an } + r{bn } = {an + rbn } is
convergent, and has limit a + rb, for all real numbers r. Thus the convergent sequences form a
linear subspace of the set of real sequences.

24. Show that Fn is the direct sum of the subspaces

W1 = {(a1 , a2 , . . . , an ) ∈ Fn : an = 0}
and
W2 = {(a1 , a2 , . . . , an ) ∈ Fn : a1 = a2 = · · · = an−1 = 0}

Solution: Given any v ∈ Fn , v = (v1 , v2 , . . . , vn ) one may express v as the sum of w1 =

(v1 , v2 , . . . , vn−1 , 0), which lies in W1 , and w2 = (0, 0, . . . , 0, vn ), which lies in W2 . Thus
Fn = W1 + W2 .
Now, suppose v = (v1 , v2 , . . . , vn ) lies in both W1 and W2 . To fit the first condition, vn = 0.
To fit the second, v1 = v2 = · · · = vn−1 = 0. Thus v1 = v2 = · · · = vn−1 = vn = 0, so v = 0.
Thus W1 ∩ W2 contains only the zero vector.
Together, Fn = W1 ⊕ W2 .

28. Prove that the set W1 of all skew-symmetric n × n matrices with entries from F is a subspace of
M× n×n (F ). Let W2 be the subspace of symmetric n × n matrices. If char(F ) 6= 2, prove that
M× n×n (F ) = W1 ⊕ W2 .

Solution: Given a matrix A ∈ Mn×n (F), define w1 = 12 (A − At ) and w2 = 12 (A + At ). Observe

that
w1t = 21 (A − At )t = 12 (At − (At )t ) = 12 (At − A) = - 12 (A − At ) = −-w1 ,
so w1 ∈ W1 , and

w2t = 21 (A + At )t = 12 (At + (At )t ) = 12 (At + A) = 12 (A + At ) = w2 ,

so w2 ∈ W2 . Finally observe that

w1 + w2 = 12 (A − At ) + 12 (A + At ) = A.

Thus Mn×n (F) = W1 + W2 .

Now suppose A ∈ W1 ∩ W2 , so A = At = -A. If char(F ) 6= 2, then A = -A ⇐⇒ Aij = -Aij
implies Aij = 0 for all ij. Thus W1 ∩ W2 contains only the zero vector.
Together, Mn×n (F) = W1 ⊕ W2 .