MSM3P22/MSM4P22

Further Complex Variable Theory & General Topology

Solutions to problem sheet 1
Jose A. Canizo
February 2013

Exercise 1.2. Let X and Y be sets and f : X Y be a function. Let A X, A P(X),

B Y , and B P(Y ). We use the shorthand
[ [
A A,
AA

and similarly for B.

A = AA f (A).
S
S
1. Prove that f

A AA f (A). Under what conditions on f do you get equality

T
T
2. Prove that f
for all collections A ? Give an example where equality does not hold.

3. Prove that f 1 B = BB f 1 (B).

S
S

4. Prove that f 1 B = BB f 1 (B).

T
T

5. Prove that A f 1 f (A) . Under what conditions do you get equality for all sets

A X? Give an example where equality does not hold.

6. Prove that f f 1 (B) B. Under what conditions do you get equality for all sets

B Y ? Give an example where equality does not hold.

7. Prove that X f 1 (B) = f 1 (Y B). What can you say about Y f (A) and f (X A)?
Under what conditions would these two sets be equal?

1. If y is in f ( A ) then there
S exist B A and x B withS
S
Solution. y = f (x). Hence
y f (B), and in particular y AA f (A). S AA f (A), then
Conversely, if y S
y f (B) for some B A . In particular, y f ( A ), since B A .

1
2. Take y f ( A ). Then y = f (x) for some x A . Since T x A for every A A ,
T T
we see that y f (A) for every A A , which implies y AA f (A).
For a counterexample to equality, take A = {1}, B = {2} and f : {1, 2} {1} given
by f (1) = f (2) = 1. Then A B = but f (A) f (B) = {1}.
T
If f is injective, then equality holds: take y AA f (A). Then y f (A) for all
A A , which means that for every A A there is xA A such that y = f (xA ).
Since f is injective, allTthe xA must be equal to one another. Call
Tthis common value
x. We have then x A and y = f (x). This shows that y f ( A ).
For an example where equality fails take X = Y = {1, 2} and f (1) = f (2) = 1. Then
f ({1} {2}) = f () = , while f ({1}) f ({2}) = {1}.
3. Using the definitions of union and inverse image we have that:
[ [
x f 1 ( B) f (x) B
f (x) B for some B B
x f 1 (B) for some B B
[
x f 1 (B).
BB

4. Using the definitions of intersection and inverse image we have that:

\ \
x f 1 ( B) f (x) B
f (x) B for all B B
x f 1 (B) for all B B
\
x f 1 (B).
BB

5. Take x A. Then f (x) f (A) (by definition of the direct image f (A)), so x
f 1 (f (A)) (by definition of the inverse image). This shows that A f 1 (f (A)).
Lets see that the reverse inclusion (and hence equality) holds if f is injective: assuming
this, take x f 1 (f (A)). Then f (x) f (A), which means that there is x0 A such
that f (x) = f (x0 ). By injectivity, x = x0 , so x A.
For an example where equality does not hold take the same f as in point 2 above and
notice that f 1 (f ({1})) = f 1 ({1}) = {1, 2} =
6 {1}.
6. Take y f (f 1 (B)). Then there is some x f 1 (B) such that f (x) = y. But
x f 1 (B) means that f (x) B, so y B.
The reverse inclusion holds if f is surjective: take y B. Since f is surjective, there
is x X such that f (x) = y B. Then x f 1 (B), and hence y f (f 1 (B)).
For an example where equality does not hold take again the same f as in point 2 above.
Then {2} =6 f (f 1 ({2})) = f () = .

2
Exercise 1.4. Let X be a topological space and let A, U and C be subsets of X. Prove the
following:

1. A is the largest open subset contained in A, A is the smallest closed set which contains
A and A A A.

2. U is open if and only if U = U and C is closed if and only if C = C.

3. If A C and C is closed, then A C.

4. A = A and A = A.

5. A = A {x X : x is a limit point of A}.

Solution. Recall that we are using the following definitions of interior and closure, the ones
given in class: [ \
int(A) := U, cl(A) := C.
U open C closed
U A CA

1. int(A) is open because it is the union of open sets, and it is contained in A because it
is the union of subsets of A. Take an open set U A. Then since U is one of the sets
whose union is taken in the definition of int(A), we have that U int A. Hence int(A)
is the largest open set contained in A.
cl(A) is closed because it is the intersection of closed sets, and it contains A because it
is the intersection of sets which contain A. (Alternatively, one can show that X \ cl(A)
is open by using De Morgans laws:
\ [
X \ cl(A) = X \ C= (X \ C),
C closed C closed
CA CA

a union of open sets). Take a closed set C A. Then since C is one of the sets whose
intersection is taken in the definition of cl(A), we have that cl(A) C. Hence cl(A) is
the smallest closed set which contains A.
The two previous remarks already show that int(A) A cl(A).

2. If int(A) = A then clearly A is open due to point 1. Conversely, if A is open then A is

one of the sets whose union is taken in the definition of int(A), so int(A) A. From
point 1 we know that int(A) A, so int(A) = A.
If cl(A) = A then clearly A is closed due to point 1. Conversely, if A is closed then A
is one of the sets whose intersection is taken in the definition of cl(A), so cl(A) A.
From point 1 we know that cl(A) A, so cl(A) = A.

3. This is a more precise way of saying that cl(A) is the smallest closed set that contains
A and was proved in point 1.

3
 4. Since int(A) is open, int(int(A)) = int(A) by point 2. Since cl(A) is closed, cl(cl(A)) =
cl(A) by point 2.

5. Let us first show that cl(A) A {x X : x is a limit point of A}. We know that
A cl(A), so we just need to show that if x is a limit point of A then x cl(A). Take
an arbitrary closed set C A. Then X \ C is open and A (X \ C) = , so we deduce
that x
/ X \ C (otherwise we would contradict the definition of a limit point for x).
In other words, x C. We have proved that x must be in every closed set C which
contains A, so x cl(A).
In order to show that cl(A) A {x X : x is a limit point of A}, take x cl(A). If
x A this is obvious, so assume x / A, and let us show that x is a limit point of A.
If U is any open set with x U , then X \ U is a closed set which does not contain x.
But since x cl(A), X \ U cannot contain A. In other words, A U 6= , and we have
proved that x is a limit point of A.

Exercise 1.6. If A, B are subsets of a topological space X, prove the following identities:

1. A B = A B.

2. A B A B.

3. int(A) int(B) int(A B)

4. int(A B) = int(A) int(B)

Give examples in which the inclusions in (b) and (c) above are strict.

Solution. 1. We have A A and B B. Hence,

A B A B.

Since A B is a closed set (being a finite union of closed sets) we have

A B A B.

On the other hand, A A B A B. Since A B is closed, this implies that

A A B.

Similarly, B A B; taken together, these imply that A B A B.

2. Since A A and B B, obviously A B A B. Since A B is closed (an

intersection of two closed sets) we have that A B A B.

3. Since int(A) A and int(B) B, obviously int(A) int(B) A B. Since int(A)

int(B) is open (a union of open sets) we have that int(A) int(B) int(A B).

4
 4. We have int(A) A and int(B) B, so int(A) int(B) A B. Since int(A) int(B)
is open (an intersection of two open sets) we have that int(A) int(B) int(A B).
On the other hand, since int(A B) A B A we have that int(A B) int(A).
Analogously, int(A B) int(B). Hence int(A B) int(A) int(B).

Exercise 1.11. Which of the following collections are bases for a topology on R?

1. B1 = {(a, b) | a, b R, a < b}

2. B2 = {[a, b] | a, b R, a < b}

3. B3 = {(a, b] | a, b R, a < b}

4. B4 = {[a, b) | a, b R, a < b}

Solution. We will use Lemma 4.4 in order to answer the question.

1. Obviously, any point x is in (x 1, x + 1), so it is contained in at least one element

of B1 . On the other hand, if we have (a, b) B1 and (c, d) B1 with nonempty
intersection, then their intersection is given by (max{a, c}, min{b, d}), which is also an
element of B1 , so point 2 of Lemma 4.4 is also satisfied. Hence B1 is the base of a
topology on R (of course, we know from Example 4.2 that B1 is indeed a base for the
usual topology of R).

2. B2 does satisfy point 1 of Lemma 4.4, but does not satisfy point 2: if we consider the
elements [1, 2] B2 and [2, 3] B2 , then [1, 2] [2, 3] = {2}, and there is no element
of B2 contained in {2}. Hence B2 is not a base of any topology on R.

3. Obviously, any point x is in (x1, x+1], so it is contained in at least one element of B3 .

On the other hand, if we have (a, b] B3 and (c, d] B3 with nonempty intersection,
then their intersection is given by (max{a, c}, min{b, d}], which is also an element of
B3 , so point 2 of Lemma 4.4 is also satisfied. Hence B3 is the base of a topology on
R. (The topology it generates is called the Sorgenfrey line; see next exercise.)

4. This is completely analogous to point 3.

Exercise 1.12. The

 Sorgenfrey line is the set R with the topology generated by the base
(a, b] : a < b R .

1. Give an example of a set A that is open and closed in the Sorgenfrey line topology but
neither open nor closed in the usual Euclidean topology on R.

2. What are the limit points, the interiors and the closures of the following subsets of R
with the Sorgenfrey
line topology?
a) (0, 1); b) [ 2, 2]; c) {q Q : 0 < q < 1}; d) {r R Q : 0 < r < 1}; e)
{1/n : 1 n N}; f) {1/n : 1 n N}; g) { 2 1/n : 1 n N}.

5
Solution. 1. (1, 2] is one such example: it is open in the Sorgenfrey topology because
it belongs to the given base, and closed in the Sorgenfrey line because R \ (1, 2] =
(, 1] (2, +) is a union of two open sets; but it is neither open nor closed in the
usual topology of R.

2. Note that the concepts of limit point, interior and closure can all be stated in terms of
a base B for the topology:

Limit point: x is a limit point of A if and only if for every B B with x B it is

true that B A 6= .
Interior: x is in the interior of A if there exists B B with x B A.
Closure: x is in the closure of A if for every B B with x B it is true that
A B 6= .

(a) (0, 1). Limit points: (0, 1]. Interior: (0, 1). Closure: (0, 1].

(b) [ 2, 2]. Limit points: [ 2, 2]. Interior: ( 2, 2]. Closure: [ 2, 2].
(c) (0, 1) Q. Limit points: (0, 1]. Interior: . Closure: (0, 1].
(d) (0, 1) (R \ Q). Limit points: (0, 1]. Interior: . Closure: (0, 1].
(e) {1/n : 1 n N}. Limit points: . Interior: . Closure: {1/n : 1 n N}.
(f) {1/n : 1 n N}. Limit points: {0}. Interior: . Closure: {1/n : 1 n
N} {0}.

(g) { 2 1/n : 1
n N}. Limit points: { 2}. Interior: . Closure: { 2 1/n :
1 n N} { 2}.