1.2 - Set Theory Review

1.2.1

a

Theorem

3 is irrational.

\begin{proof}
Assume for contradiction that 3 is rational, so then:

pq=3

for some p,q∈Z where q≠0 and p,q don't share any common factors. Squaring both sides gives that:

p2=3q2

so 3|p2, or p2 has a factor of 3. Notice that because of that:

p23=p3⋅p∈Z

so then since p∈Z then clearly 3|p so then set p=3k for some k∈Z so then:

(3k)2=3q2⇒9k2=q2⇒q2=3p2

so then q2 is divisible by 3, and by the same argument 3|q which contradicts p,q not sharing any factors.
\end{proof}
Note that the argument would follow similarly for 6, except that when we breakup p2/6 notice that 6 isn't prime so then we'll get:

p26=p2⋅p3

so we show that p is either divisible by 2 or by 3 and continue the argument to show that q shares the same factor.

b

The proof for showing 4 would work in the same way it did for 1.2 - Set Theory Review#1.2.1#a since we'd get to:

p2=4q2=2(2q2)

so then p is even, so then p=2k:

(2k)2=2(2q2)⇒4k2=4q2⇒k2=q2

but this couldn't lead to a contradiction as k=|q| so then p=2|q| so then:

pq=2|q|q=±2

which doesn't give a contradiction.

1.2.3

a

Postulate

If A1⊇… are all sets with an infinite number of elements, then ⋂n=1∞An is infinite as well.

This is a false statement. Consider the superset (0,1)⊆R, and have Ai defined by:

Ai=(0,12i)

We can see that Ai+1=Ai∩(12i+1,1) so then by that definition then Ai+1⊇Ai as required. But all of these sets are uncountably infinite, but notice that ⋂n=1∞An=∅. This is because if there were some x∈⋂n=1∞An then I can construct a new set A′ where:

A′=(0,x2)

where clearly A′ is in our sequence of sets, but x∉A′ by definition. But this larger set is finite in length (length 0) which is a counter example to the postulate at hand.

b

Postulate

If A1⊇… are all sets with finite number of elements, then ⋂n=1∞An is finite as well.

This is a true statement. Namely, we know that if A1⊇A2 then A1∩A2=A1 which is finite, and in general then A1⊇⋯=A1 so then clearly ⋂n=1∞An=A1 is finite.

c

Postulate

A∩(B∪C)=(A∩B)∪C

This is a false statement. Let A={1,2},B={2,3} and C={3,1}. Notice that:

A∩(B∪C)={1,2}∩{1,2,3}={1,2}

but:

(A∩B)∪C={2}∪{3,1}={1,2,3}

and these don't equal.

d

Postulate

A∩(B∩C)=(A∩B)∩C

This is a true statement. For ⊇ let x∈(A∩B)∩C so then x∈C and x∈A∩B so x∈A and x∈B. Then x∈B∩C and thus x∈A∩(B∩C). The other direction is proved similarly WLOG.

e

Postulate

A∩(B∪C)=(A∩B)∪(A∩C)

This is a true statement.

⊆. Let x∈A∩(B∪C). Then x∈A and x∈B∪C. If x∈B then x∈A∩B so then x∈(A∩B)∪(A∩C). Then if x∉B then x∈C by requirement so then similarly x∈A∩C and x is in the RHS set. Thus we get ⊆.

⊇. Let x be in the RHS set. Suppose x∈A∩B so then x∈A and x∈B. Then x∈B∪C so then x∈A∩(B∪C). WLOG the same applies for C.

1.2.5 (De Morgan's Laws)

a

Theorem

(A∩B)c⊆Ac∪Bc

\begin{proof}
Let x∈(A∩B)c. Then x∉A∩B. For one case if x∈A then we must have that x∉B so then x∈Bc so then x∈Ac∪Bc. If instead x∉A then x∈Ac and we get the same result.
\end{proof}

b

Theorem

(A∩B)c⊇Ac∪Bc

\begin{proof}
Let x∈Ac∪Bc. If x∈Ac then x∉A, so clearly x∉A∩B. Thus, then x∈(A∩B)c. WLOG, the same applies if x∈Bc.
\end{proof}

c

Theorem

(A∪B)c=Ac∩Bc

\begin{proof}
First ⊆. Let x∈(A∪B)c. Then x∉A∪B. Notice if x∈A then that's a contradiction and same for x∈B, so then x∉A and x∉B so then x∈Ac∩Bc.

Then ⊇. Let x∈Ac∩Bc, thus x∈Ac and x∈Bc. Then x∉A and x∉B. Notice by similar reasoning prior that x∉A∪B so x∈(A∪B)c.
\end{proof}

1.2.7

Given function f and a subset of its domain A, we denote f(A)={f(x):x∈A} as the range of f over set A.

a

Given f(x)=x2 and A=[0,2] and B=[1,4] then we know that:

f(A)=[0,4],f(B)=[1,16]

so then:

f(A∩B)=f([1,2])=[1,4]

while:

f(A)∩f(B)=[1,4]

so in this case f(A∩B)=f(A)∩f(B). Similarly

f(A∪B)=f([0,4])=[0,16]

while:

f(A)∪f(B)=[0,16]

so f(A)∪f(B)=f(A∪B) in this case.

b

In general we don't have f(A∩B)=f(A)∩f(B). For instance, if A=[−1,0] and B=[0,2] then:

f(A)=[0,1],f(B)=[0,4]f(A∩B)=f([0])={0}f(A)∩f(B)=[0,1]

here clearly f(A∩B)≠f(A)∩f(B).

c

Theorem

For any g:R→R then g(A∩B)⊆g(A)∩g(B).

\begin{proof}
Let y∈g(A∩B) be arbitrary. Then there is some x∈A∩B such that f(x)=y by definition. Clearly x∈A and x∈B. Thus, since f(x)=y then y∈f(A) since x∈A and likewise y∈f(B). Thus, y∈g(A)∩g(B), completing the proof.
\end{proof}

1.2.9

Given some f:D→R and a subset B⊆R, denote f−1 as the preimage, namely:

f−1(B)={x∈D:f(x)∈B}

a

Let f(x)=x2. If A is the closed interval [0,4] and B is the closed interval [−1,1] then:

f−1(A)=[−2,2],f−1(B)=[−1,1]

Notice that:

f−1(A∩B)=f−1([0,1])=[−1,1]

while:

f−1(A)∩f−1(B)=[−1,1]

so in this case f−1(A∩B)=f−1(A)∩f−1(B). Similarly:

f−1(A∪B)=f−1([−1,4])=[−2,2]

while:

f−1(A)∪f−1(B)=[−2,2]

so then f−1(A∪B)=f−1(A)∪f−1(B) in this case.

b

Theorem

For any g:R→R then:

  • g−1(A∩B)=g−1(A)∩g−1(B)
  • g−1(A∪B)=g−1(A)∪g−1(B)
    for all sets A,B⊆R.

\begin{proof}
We prove the former.

(⊆). Let x∈g−1(A∩B). Then there is some f(x)∈A∩B by definition. Then f(x)∈A and f(x)∈B separately. Thus then x∈g−1(A) and similar for B, so then x∈g−1(A)∩g−1(B).

(⊇). Let x∈g−1(A)∩g−1(B). Then x∈g−1(A) and similar for B. Then there is some f(x)∈A and likewise f(x)∈B. Thus f(x)∈A∩B. Thus, x∈g−1(A∩B).

For the latter:

(⊆). Let x∈g−1(A∪B). Then there's some f(x)∈A∪B. If f(x)∈A then x∈g−1(A) so then x∈g−1(A)∪g−1(B). WLOG, the same can be sead if f(x)∈B instead (ie: f(x)∉A).

(⊇). Let x∈g−1(A)∪g−1(B). If x∈g−1(A) then f(x)∈A so then f(x)∈A∪B so then x∈g−1(A∪B). WLOG, the same works if x∈g−1(B) instead.
\end{proof}

1.2.11

We'll negate certain statements into their opposites:

a

The opposite of "for all reals where a<b then there is some n∈N where a+1/n<b" is:

"there are reals a<b where, for any n∈N, we have it that a+1/n≥b".

Here the original claim is true, as we can always choose some lowest fraction that's less than some ε>0.

b

The opposite of "there is a real x>0 such that x<1/n for all n∈N" is:

" for all reals x>0 there is some n∈N such that x≥1/n"

Here the negation is true for similar reasons to 1.2 - Set Theory Review#1.2.11#a.

c

The opposite of "between every two distinct real numbers is a rational number" is:

"there are two real numbers such that there exist no rational numbers between them"

I'm pretty sure that the negation is true since Q is not complete.

1.2.13

We'll be allowed to use De Morgan's Laws here from 1.2 - Set Theory Review#1.2.5 (De Morgan's Laws).

a

Theorem

For any n∈N:

(A1∪A2∪⋯∪An)C=A1c∩⋯∩Anc

\begin{proof}
We prove this via induction. The base case for A1,A2 is just the simpler De Morgan's Law from 1.2 - Set Theory Review#1.2.5 (De Morgan's Laws)#c. Let k∈N be arbitrary, and suppose the theorem holds for n≤k. Consider the k+1 case, then:

(A1∪⋯∪Ak∪Ak+1)c=([A1∪⋯∪Ak]∪Ak+1)c

Since the LHS set is composed of n=k≤k sets and the RHS set is just one set, we can use our Inductive Hypothesis:

=[A1∪⋯∪Ak]c∩Ak+1c=A1c∩⋯∩Akc∩Ak+1c

completing the proof by induction.
\end{proof}

b

We'll show that induction can't always be used. We'll show that ⋂i=1nBi≠∅ for some n∈N but ⋂i=1∞Bi=∅. Let Bi=(0,12i). Here any finite intersection allows us to find some x in that intersection, but intersecting all infinitely many sets, as we saw in 1.2 - Set Theory Review#1.2.3#a that we get the empty set.

c

Theorem

(⋃i=1∞Ai)c=⋂i=1∞AiC

\begin{proof}
We prove (⊆) first. Let x be an element of the LHS. Then x∉⋃i=1∞Ai. That means that there isn't any set Ai where x∈Ai. Hence, for all sets Ai then x∉Ai, so then x∈Aic for all these sets. Hence, then x∈⋂i=1∞Aic as required.

Now for (⊇). Let x be an element of the RHS. Then x∈Aic for all i. Hence, then x∉Ai for all i. Then there doesn't exist any sets Ai where x∈Ai so then x∉⋃i=1∞Ai. Hence, x is in the LHS.
\end{proof}