Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Introduction to Abstract Mathematics

MA 103

Solutions to exercises 4

1 (a) There are several examples you could use. You saw in lectures that S = {True,False}
and ∗ is =⇒ works: (False implies True) is True; (True implies False) is False. Another
example is subtraction on Z: 3 − 5 is not equal to 5 − 3. Important note: subtraction on
N is not a good example, because subtraction is not a binary operation on N. This is
because the answer might not be a member of N: 3 − 5 is not in N.
(b) Again there are several examples. You worked out in Exercises 1, Q4, that the same logic
example works: =⇒ is not associative. Another possibility is division: ( a/b)/c is not
generally equal to a/(b/c), for example (2/2)/2 = 21 , but 2/(2/2) = 2. But you need
to be careful here to make sure you write down a binary operation. So you cannot use
division on Z, because 1/2 is not in Z. You also cannot use division on R, because 1/0
is not defined and a binary operation has to be defined for all pairs of elements. What
does work is division on R \ {0}. You can quickly check that if you take any two real
numbers which are not zero and divide one by the other, you get a real number which
is not zero.

2 A mistake many people made in this question was either to do subtraction or to try to use
0. 0 is not a natural number, so you cannot add it—the axioms don’t apply to things which
are not natural numbers. And we did not define subtraction yet (because it’s not a binary
operation on N, as we saw in (1) ).
(a) For each natural number n, there is no natural number strictly between n and n + 1. In
logical notation this is

∀ n ∈ N : ¬ ∃ q ∈ N : ( n < q ) ∧ ( q < n + 1) .


You might shorten the (n < q) ∧ (q < n + 1) by writing n < q < n + 1. There are also
other valid ways to write this.
To prove it, you should certainly look back to the lecture slides where we proved the
special case n = 1. If you do this, you will see that exactly the same method works:
Proof Given n ∈ N, suppose for a contradiction that there exists q ∈ N such that
n < q < n + 1.
By Axiom 11, we have q = n + x for some x ∈ N. Since q < n + 1, by Axiom 12 we have
q 6= n + 1, so by definition we have

n + x = q 6= n + 1 ,


c London School of Economics, 2018
MA 103 Introduction to Abstract Mathematics Solutions to exercises 4 — Page 2

and by Axiom 2 (commutativity) we get

x + n 6= 1 + n .

Now (if x = 1 then x + n = 1 + n) is True by definition. The contrapositive of this


statement is: (if x + n 6= 1 + n then x 6= 1). Since we have x + n 6= 1 + n, we conclude
x 6= 1.
Since 1 is the smallest natural number, as we proved in lectures, by Axiom 12 we have
1 < x.
As we proved in lectures, since 1 < x we have 1 + n < x + n, which by Axiom 2 is
equivalent to n + 1 < n + x. Recall that we also have q = n + x, so we now proved
n + 1 < q.
But we assumed to start with that n < q < n + 1, which in particular means n + 1 > q.
Axiom 12 doesn’t allow both n + 1 > q and n + 1 < q to be true, so we have arrived at
a contradiction.
It follows that our original assumption must be false, i.e. there exists no such q. That’s
what we wanted to prove.
There are other ways to prove this statement. It’s natural to think of trying induction.
But you should not have done this because I told you that you should only try induction
if you get stuck trying a proof another way—and you should not have got stuck!
It turns out to be rather more painful to use induction to prove this statement. You can,
but you end up doing most of the work in the direct proof and having to wrap it all up
in an induction.
(b) For all natural numbers a and b such that a < b, there is no natural number c such that
b + c = a.
In logical notation we can write

∀ a, b ∈ N : a < b =⇒ ¬(∃c ∈ N : b + c = a) ,


or

∀ a, b ∈ N with a < b : ¬ ∃c ∈ N : b + c = a .


This is an immediate consequence of Axioms 11 and 12.


Proof Given a, b ∈ N such that a < b, suppose there exists c ∈ N such that b + c = a.
By Axiom 11, since b + c = a we have b < a. But Axiom 12 says we cannot have both
a < b and b < a, so this is a contradiction.
It’s worth checking you understand why the proof goes like this. The statement we
want to prove is that something holds for all a, b ∈ N such that a < b. In other words, if
you give me any a, b ∈ N such that a < b, then I should be able to prove that for these
specific a and b the ‘something’ is true. This is why the first line of the proof is ‘Given
a, b ∈ N such that a < b...’. This is the usual way we prove ‘for all...’ statements.
It’s also worth commenting on what this proof says. In lectures (Thursday week 4) I
said that the natural numbers are 1, 2, 3, . . . and nothing else. But why are these numbers
really all different? Maybe you count 1, 2, 3, . . . , 24, 1 and loop back to where you started,
MA 103 Introduction to Abstract Mathematics Solutions to exercises 4 — Page 3

or to some earlier point? Of course you know that this doesn’t really happen in N. (But
it does happen in for example Z24 , and in ‘finite groups’ (to be defined next term) in
general)
Well, this question proves that that doesn’t happen. If you looped back, then you would
add one to some number b and get to an earlier number a, i.e. we would have b + 1 = a
where a < b. We just proved this is not possible.

3 (a) This question is really just a case of noticing one trick, which we saw in the lectures
(when we talked about what amounts of money we can pay with 4 and 5 pound coins).
But it is written, deliberately, in an abstract way which makes it hard to see when you
are not confident with working with abstract logic.
Let’s first give the (very short) answer.
The statement we are trying to prove is the Principle of Induction with base case N.
We prove it as follows. Define a predicate Q by Q(m) ⇐⇒ P( N + m − 1) and then
what we want to prove is exactly the Principle of Induction for Q, which we already
know is true.
That’s the answer. If you are happy with the abstract setting and the logic, then you will
probably be happy that this was enough to answer the question — if you’re revising for
the exam when you read this, hopefully that was the case.
But if you are not very happy with the abstract setting and the logic, your reaction to
that answer is probably ‘what are you talking about’. So let’s go through it a bit more
slowly and explain what’s going on. To get started you need to read a logical statement
and understand it. That means you need to break it up into pieces and figure out what
the pieces mean.
We are supposed to prove that something is true for all integers N and predicates P(t).
So our proof is going to start ‘Given N and P(t)...’ and these things are now fixed. N is
an integer, and P(t) is a logical statement which can be either true or false, depending on
the integer t. It’s important to remember that ‘P(t)’ is simply a name for a string of sym-
bols (which presumably includes the letter t) which is a logical formula that evaluates
to True or to False.
If you find this hard to follow, it might help you to fix a concrete example: take N = 4
and P(t) to be the statement ‘t2 ≤ 2t ’ (We did this example in the lectures in week 3).
Check that what we are going to do in the abstract setting makes sense for this concrete
example.
What we want to prove is:
  
P( N ) ∧ ∀n ∈ { N, N + 1, . . . } : P(n) =⇒ P(n + 1) =⇒ ∀n ∈ { N, N + 1, . . . } : P(n) .

Remember N and P are fixed things in this statement. There is also an n which appears
inside quantifiers. In fact, there are two different ns, because a quantifier ‘for all n...’
appears twice in the statement. This means that the two different ns have nothing to do
with each other, but we confusingly use the same letter each time. So to avoid confusion,
MA 103 Introduction to Abstract Mathematics Solutions to exercises 4 — Page 4

let’s change the letter. What we want to prove is exactly the same statement as:
  
P( N ) ∧ ∀s ∈ { N, N + 1, . . . } : P(s) =⇒ P(s + 1) =⇒ ∀t ∈ { N, N + 1, . . . } : P(t) .

Now we avoided confusion. Let’s try to decode the statement. It is of the form A =⇒
B, where both A and B are complicated statements. Let’s first look at the right hand side
(the B) because this is simpler.
The conclusion of this implication we are trying to prove is that P( N ), P( N + 1), P( N +
2) and so on are all true statements.
At this point you maybe should notice that this is the same conclusion as the Principle
of Induction with base case N. So maybe that is the statement we are trying to prove.
The premise of the implication we are trying to prove is that P( N ) is true, and all of
P( N ) =⇒ P( N + 1), and P( N + 1) =⇒ P( N + 2), and P( N + 2) =⇒ P( N + 3), and
so on, are true statements. That’s indeed what the Principle of Induction with base case
N has as a premise. Let’s repeat this.
The Principle of Induction with base case N says that if
P( N ), and
P( N ) =⇒ P( N + 1), and
P( N + 1) =⇒ P( N + 2), and so on, are all True, then P( N ), P( N + 1), P( N + 2) and so
on are all True.
And we figured out that this is exactly what the string of logical symbols also says. So
now we know what we want to prove.

What we are supposed to use is the normal Principle of Induction with base case 1 (the
one we proved works in lectures). The trick is simple:
Define a new predicate Q(m) for m ∈ N, by Q(m) is True if and only if P(m + N − 1)
is True. So Q(1) is the same statement as P( N ), and Q(2) is the same statement as
P( N + 1), and so on.
Let’s recall that what we are trying to prove,
  
P( N ) ∧ ∀s ∈ { N, N + 1, . . . } : P(s) =⇒ P(s + 1) =⇒ ∀t ∈ { N, N + 1, . . . } : P(t) ,

is a statement of the form A =⇒ B. Remember that to prove such a statement, we start


by assuming A and then we do some work until we deduce B; that proves A =⇒ B is
True.
Proof We assume
 
P( N ) ∧ ∀s ∈ { N, N + 1, . . . } : P(s) =⇒ P(s + 1) .

In other words, we assume P( N ) is true, and P( N ) =⇒ P( N + 1), and P( N + 1) =⇒


P( N + 2), and so on are all True.
Now Q(1) is exactly the same as P( N ) by definition, so because P( N ) is true, so is Q(1).
Because Q(1) and Q(2) are exactly the same as P( N ) and P( N + 1), the statement
P( N ) =⇒ P( N + 1) is exactly the same as Q(1) =⇒ Q(2). We assumed that
P( N ) =⇒ P( N + 1) is True, so Q(1) =⇒ Q(2) is also True.
MA 103 Introduction to Abstract Mathematics Solutions to exercises 4 — Page 5

And this pattern continues: from our assumption, we conclude that


Q (1)
Q(1) =⇒ Q(2)
Q(2) =⇒ Q(3)
and so on, are all True.
But this is the assumption of the normal Principle of Induction with base case 1. So by
the Principle of Induction, we conclude that Q(1), Q(2), Q(3) and so on are all True.
Now Q(m) is exactly the same as P( N + m − 1). So what this last line says is exactly the
same as saying that P( N ), P( N + 1), P( N + 2) and so on are all True.
That is what we wanted to prove.
What we just did is prove that the Principle of Induction with changed base case works.
In other words, we proved using formal abstract logic that a method of doing proofs
works. It won’t get more abstract than this in MA103; if you found it hard (you probably
did) then it is because you’re having to do something much more abstract than anything
you ever did before.
We actually saw this trick before in lectures, when we proved that with £4 and £5 coins
we can pay £12, £13, £14, £15, and so on. There we let P(1) be the property that we can
pay £12, and so on—it’s exactly the same idea, just not written in a very abstract way.
(b) P( N ) is the same as P(9), since N = 9, and P( N + 1) is the same as P(10). Now P(9)
says ‘92 ≥ 8 × 9 + 15’, i.e. 81 ≥ 87. This is False. And P(10) says 102 ≥ 8 × 10 + 15,
i.e. 100 ≥ 95. This is True. Since (False implies True) is a True statement, P( N ) =⇒
P( N + 1) is True. But P( N ) is False.
If you fell into a trap of writing some induction-like proof, next time be careful to read
the question and not be fooled into writing an answer which looks like a pattern you
expect.
(c) Again, we saw a bit of the idea in lectures. There are two ways to do this; they are about
equally good. I’ll give both. The first is to modify the proof that induction works.
Proof Suppose for a contradiction that P(1) is true, and P(1) =⇒ P(2) is true, and
P(1) ∧ P(2) =⇒ P(3) is true, and so on, but that ∀n ∈ N : P(n) is not true.


What it means that ∀n ∈ N : P(n) is not true is that the set S = {n ∈ N : ¬ P(n)}
is non-empty (this is saying the same thing in set-theoretic language). But then S is a
non-empty set of natural numbers, and Axiom 13 says that it has a least element s.
Now because s ∈ S, the statement P(s) is false. But because s ∈ S is the least element of
S, if t is a natural number such that t < s we know t 6∈ S. By definition of S, that means
P(t) is True. What we just proved is that for all t ∈ N with t < s the statement P(t) is
True. In other words, P(1) and P(2) and ... and P(s − 1) are all True. So P(1) ∧ P(2) ∧

· · · ∧ P(s − 1) is True.

But we know that P(1) ∧ P(2) ∧ · · · ∧ P(s − 1) =⇒ P(s) is True—that’s part of
what we assumed at the start of this proof. It follows that P(s) is True. But that is a
contradiction—we assumed s ∈ S, which means P(s) is False.
Another way to do it is to use the normal Principle of Induction. To do that you have to
define a clever new predicate.
MA 103 Introduction to Abstract Mathematics Solutions to exercises 4 — Page 6


Proof Suppose that True =⇒ P(1), and P(1) =⇒ P(2), and P(1) ∧ P(2) =⇒ P(3),
and so on, are all True.
We define a predicate Q(m) to mean ∀t ∈ N : t < n =⇒ P(t) .


So Q(1) is a vacuously true statement (there is no t < 1 in the natural numbers!). Q(2)
is the statement P(1). Q(3) is the statement P(1) ∧ P(2). Q(4) is the statement P(1) ∧
P(2) ∧ P(3), and so on.
Let’s try to prove that Q(n) is true for every natural number n by induction. To begin
with, the base case Q(1) is trivial: we just saw that Q(1) is True.
Now we need to prove that for a given k ∈ N we have Q(k ) =⇒ Q(k + 1).
To do that, we assume Q(k ) and aim to prove Q(k + 1) (just as we would to prove any
statement of the form A =⇒ B).
So let’s assume Q(k ) is true, i.e. P(1) ∧ P(2) ∧ · · · ∧ P(k − 1) is True.

We know that P(1) ∧ P(2) ∧ · · · ∧ P(k − 1) =⇒ P(k ) is a True statement, because
that’s part of what we assumed at the start of the proof. So P(k ) is a true statement.

But that means P(1) ∧ P(2) ∧ · · · ∧ P(k − 1) ∧ P(k ) is a True statement, and that state-
ment is exactly Q(k + 1).
So, assuming Q(k), we proved Q(k + 1); in other words, we proved that Q(k ) =⇒
Q(k + 1) is True. Since k was an arbitrary element of N, we proved Q(k ) =⇒ Q(k + 1)
is True for all k ∈ N.
By the Principle of Induction, Q(n) is True for every n ∈ N.
Now given any m ∈ N, we see that P(m) is True, because Q(m + 1) is True and Q(m + 1)
can only be True if P(m) is True (because Q(m + 1) is a collection of statements ‘and’ed
together, and one of those statements is P(m) ). So we proved P(m) is True for all m ∈ N,
which is what we wanted.
(d) We can prove this by putting the idea for (a) together with Strong Induction. Given the
predicate P(n) and N, let for each n ∈ N the predicate Q(n) be ‘P( N + n − 1) is True’.
Then the statements we are given, in terms of Q, are that the following are all True:
Q (1)
Q(1) =⇒ Q(2)

Q(1 ∧ Q(2) =⇒ Q(3)

Q(1 ∧ Q(2) ∧ Q(3) =⇒ Q(4)
...
and Strong Induction tells us that given all these statements are True, we conclude that
Q(n) is True for all n ∈ N.
Rephrasing this in terms of P, we see that P( N + n − 1) is True for all n ∈ N, which is
what we wanted to prove.

4 For un , we start with u1 = 1 and add 3 each time, so we have un = 1 + 3(n − 1) = 3n − 2.


MA 103 Introduction to Abstract Mathematics Solutions to exercises 4 — Page 7

For vn , here are the first few values without doing the arithmetic:

v1 = 1,
v2 = 22 ,
v3 = 32 × 22 ,
v4 = 42 × 32 × 22 .

We can alternatively write, for instance, v4 = (4 × 3 × 2 × 1)2 = (4!)2 . By now, it should be


apparent that vn = (n!)2 .
For the last one, we can write simply:
n n
3
wn = ∑ 3i = 3 ∑ i = 2
n ( n + 1).
i =1 i =1

It’s not hard to write down formal induction proofs for each of these formulae. Make sure
that you remember to check the base case!

5 Let x be the ratio of the rectangle, and suppose this is the length of its longer side and 1 the
length of its shorter side. After halving it the longer side has length 1 and the shorter side
has length x/2. So we require that x/1 = 1/( x/2) or x = 2/x, that is, x2 = 2 and thus

α = x = 2.
In order to determine the dimensions of A0,plet its lengths be L (for the longer side) and
√ √ √ √
L/ 2. Then L · L/ 2 = 1 and thus L = 2 = 4 2 ≈ 1.189297... metres. We have
the following dimensions, rounded to the nearest millimetre: A0: 1189 mm × 841 mm, A1:
841 mm × 595 mm, A2: 595 mm × 420 mm, A3: 420 mm × 297 mm, A4: 297 mm × 210 mm.
Note that if you started with the A4 line rounded, and simply multiplied up, you would get
the wrong answers for A2, A1 and A0. This is because you would make a rounding error:
A4 is not exactly 297 by 210 millimetres.
A0 is one square metre, A1 is half a square metre, etc., in general Ak is 1/2k of a square metre,
so A4 is 1/16 of a square metre. So at 80g/sqm the A4 sheet weighs 80/16 grams, that is, 5
grams.
MA 103 Introduction to Abstract Mathematics Solutions to exercises 4 — Page 8

6 (a) First off, the n = 1 case of the formula is true: a1 = 1 = 20181−1 .


Now, we are given an = 1 + 2017 × ∑in=−11 ai for n ≥ 2. So in order to find out what an
is—which we need to do if we want to check it’s equal to 2018n−1 —we need to know a1 ,
a2 , a3 and so on up to an−1 .
In other words, we need to know our desired formula is true for all cases from 1 up to
n − 1. That is exactly the setup for strong induction. Let’s do it formally.
Proof Define P(n) to be the statement ‘an = 2018n−1 ’. We need to prove True =⇒ P(1),

and P(1) =⇒ P(2), and P(1) ∧ P(2) =⇒ P(3), and so on.
The first of these is easy. True =⇒ P(1), as we saw, is logically equivalent to P(1). And
P(1) is True; the formula does hold for a1 .
We still need to prove, for all t ∈ N with t ≥ 2, that

P (1) ∧ P (2) ∧ · · · ∧ P ( t − 1) =⇒ P(t)

is true.
Given t ∈ N with t ≥ 2, let’s prove it. Because it is a statement of the form A =⇒ B, to
prove it we assume A and aim to deduce B
So assume P(1), and P(2), and P(3), and so on up to P(t − 1), are True.
−1
Now at = 1 + 2017 × ∑it= 1 ai by definition. Because P (i ) is true for each integer 1 ≤ i ≤
t − 1 (this is what we just assumed, the induction hypothesis) we know
t −1
at = 1 + 2017 × ∑ 2018i−1
i =1

−1 t −1
and we know 1 + 2018 + 20182 + · · · + 2018t−2 = 2018
2018−1 , because we saw this formula
several times in lectures and also proved it last week (Exercise 2).
So we have
2018t−1 − 1
at = 1 + 2017 × = 1 + 2018t−1 − 1 = 2018t−1 .
2018 − 1
This is the statement P(t). So assuming P(1), and P(2), and P(3), and so on up to
P(t − 1), are True, we proved P(t). In other words, we proved

P (1) ∧ P (2) ∧ · · · ∧ P ( t − 1) =⇒ P(t) .

Our proof works for each t ∈ N with t ≥ 2—we didn’t make any assumption about
what t is other than that it is in N and at least 2—so we proved

P (1) ∧ P (2) ∧ · · · ∧ P ( t − 1) =⇒ P(t)

is True for all t ∈ N with t ≥ 2. Remember we already proved the t = 1 case. So by


Strong Induction, we conclude that P(n) is true for all n ∈ N, i.e. that an = 2018n−1 for
all n ∈ N. That’s what we wanted to prove.
You might ask whether we needed to prove the t = 1 case separately in the above. The
answer is we did not. Our t ≥ 2 proof works as well for t = 1—but to see this you have
to be a bit careful that you understand correctly ∑0i=1 ai = 0. This is true—that’s how we
MA 103 Introduction to Abstract Mathematics Solutions to exercises 4 — Page 9

defined it in lectures, and we did so precisely because it makes things work more easily
in problems like this. So we could have skipped that—but if you’re not sure, it’s always
best to be careful.
You could also wonder whether you really need to write out all this detail. You do not;
in the exam something more like the following would be fine.
Proof We prove it using Strong Induction. Let P(n) be the statement an = 2018n−1 . For
n ≥ 1, assume P(k ) holds for each 1 ≤ k ≤ n − 1 as an induction hypothesis. Then we
have
n −1
2018n−1 − 1
an = 1 + 2017 × ∑ 2018i−1 = 1 + 2017 × 2018 − 1
= 1 + 2018n−1 − 1 = 2018n−1
i =1

which is P(n), so we proved P(1) ∧ · · · ∧ P(n − 1) =⇒ P(n). By Strong Induction, we


see P(n) holds for all n ∈ N.
Everything is here—in particular, we are being careful to say what we prove when we
show the induction step works: we prove an implication is true. We do NOT prove P(n)
is true until we say ‘by Strong Induction..’.
Finally, if you are clever you can avoid Strong Induction here, by using the recursion
formula both for an and an−1 in order to find a recursion for an in terms only of an−1 .
However, if you do this you either need to point out that the recursion given for n ≥ 2
also works for n = 1, or alternatively you need to have base cases a1 and a2 ; this trick
only works for n ≥ 3.
(b) The first part of this is really just asking you to solve simultaneous linear equations:
√  √ 
1+ 5 1 1− 5 1
2=A 2 +B 2 , and
√  √ 
1+ 5 2 1− 5 2
11 = A 2 +B 2 .

You learnt how to do this in school: write A in terms of B, substitute, solve for B, put it
back into your formula for A and check the answer. You should get
√ √
9 √5−5 5+√
9 5
A= 2 5
and B= 2 5
.

And you should check it. This is why I asked you to check your formula gives the right
answer for b3 , in order to make sure you really check.
Now we need to show the formula works for all n, which of course we will do by in-
duction. Since we need to know the formula holds for bn−1 and bn−2 in order to prove it
for bn , we’ll use strong induction.
For n = 1, 2 we already showed the formula holds. Given n ≥ 3, assume as an induction
hypothesis that
√  √ 
1+ 5 t 1− 5 t
bt = A 2 +B 2

holds for each 1 ≤ t ≤ n − 1.


MA 103 Introduction to Abstract Mathematics Solutions to exercises 4 — Page 10

We have
√  √  √  √ 
n −1 n −1 n −2 1− 5 n −2
bn = bn−1 + bn−2 = A 1+2 5 + B 1−2 5 + A 1+2 5 +B 2
√   √ √   √
n −2 1+ 5 n −2 1− 5
 
= A 1+2 5 2 + 1 + B 1−2 5 2 +1
√  √ √  √
n −2 1+2 5+5 1− 5 n −2 1−2 5+5
   
= A 1+2 5 4 + B 2 4
√  √  √  √ 
n −2 1+ 5 2 n −2 1− 5 2
= A 1+2 5 2 + B 1−2 5 2
√  √ 
1+ 5 n 1− 5 n
=A 2 +B 2 .

This is the formula we wanted to get to. So for each n ≥ 3 our induction hypothesis
above implies
√  √ 
n n
bn = A 1+2 5 + B 1−2 5 ,

which completes the induction step. By Strong Induction we conclude that


√  √ 
n n
bn = A 1+2 5 + B 1−2 5

holds for every n ∈ N.

Again, there are two important things to note here. First, we didn’t formally define a
predicate; we just wrote down the statement we’re trying to prove. We could have been
more formal and defined P(t) to be
√  √ 
t t
bt = A 1+2 5 + B 1−2 5 ,

but we don’t have to do this as long as we are clear what we are doing.
Second, we need to be clear about what we proved at what point. On the line where we
say ‘completes the induction step’, what we just proved is an implication: if our formula
holds for all bt with 1 ≤ t ≤ n − 1 then it also holds for bn . At this point we have NOT
proved that the formula holds for bn !
We have only proved that the formula holds for bn once we use Strong Induction.
This is an important difference for the following reason. A standard error in induction
proofs is to forget to do the base case, only prove the induction step, and then say ‘we
proved it for all n’. This is not true. If you think of an induction proof as like climbing
a ladder, from rung 1 to 2 to 3 and so on, the induction step says if you have got to
rung n − 1 then you can take the next step up to rung n. If the first rung—the n = 1
base case—is missing, you still cannot climb the ladder. And in fact, it’s easy to ‘prove’
wrong formulae by forgetting to prove the base case.
Try to prove ∑in=1 i = 100 + 21 n(n + 1) by induction. You’ll find the induction step works
perfectly. But the base case is false: 1 is not equal to 101. And the formula is false for
all n. It is never true. When you proved the induction step, you only proved that if the
formula holds for n then it holds for n + 1. You did NOT prove that the formula holds
for n + 1. And in fact it doesn’t.

You might wonder: how on earth did it work that we did not need to know what A
and B are in order to do the induction step? (In fact, you probably wrote out A and B
explicitly, and it made the algebra much harder).
MA 103 Introduction to Abstract Mathematics Solutions to exercises 4 — Page 11

Well, it turns out that if c1 and c2 are given, and cn = cn−1 + cn−2 for every n ≥ 3, then
we always have a formula
√  √ 
1+ 5 n 1− 5 n
cn = α 2 +β 2

for some α and β which we can determine given c1 and c2 . That’s more or less exactly
what we just proved. Why, and how does this work for other recursions? Come to
MA210 next year and find out.

7 (a) This part should be easy to answer. If S = {Ø}, then S+ = {Ø, {Ø}}. And (S+ )+ =
{Ø, {Ø}, {Ø, {Ø}}}. It probably looks horrible, but you can see the pattern.
The rest of this question is about something called the ‘von Neumann construction’, which
is a way to start from set theory and construct a set which behaves exactly like N (so by
Dedekind’s Theorem it is isomorphic to N). This sounds like it answers the problem I men-
tioned (as non-examinable philosophy) in lectures: what if N doesn’t exist? But of course,
what we have done is not get rid of the problem, but push it somewhere else. The von
Neumann construction only exists if you assume set theory exists. Which we cannot prove
(except by assuming some other complicated structure exists, which is again just pushing
the problem around).
In some textbooks, particularly on the foundations of mathematics, you may see it said that
{Ø} is 1, and {Ø, {Ø}} is 2, and so on. I strongly suggest you ignore this: 1 is 1 is a number,
it is not some funny set. You can define 1 to be a funny set; it will work. But it is very easy
to get yourself confused, so don’t do it.
If you go and look up the details, you will see that usually the von Neumann construction
starts with Ø, which is supposed to correspond to 0. I’ve omitted Ø and started with {Ø}
in order that we get a copy of N (and not N ∪ {0} ). This turns out not to be an important
difference.
What you also see (sort of—many details are missing) in this question is that we don’t really
need to assume all the axioms of N we wrote down. You could get away with a few axioms
that talk about how this a+ operation, called ‘successor’ behaves (it is basically just ‘add
one’). And then you can define addition, multiplication and < and you will get all the
axioms we wrote down. This is called Peano arithmetic; the axioms are the ‘second-order
Peano axioms’ (look them up if you want to know more). This is the standard way to give
axioms for N. The reason is that with fewer simpler axioms, it’s easier to check that a given
structure satisfies them (or doesn’t). But the Peano axioms are not anything like other axioms
you’ll see later in this course (for groups and vector spaces and so on) so I did not talk about
them in the lectures. It is also hard work to obtain ‘useful’ statements from the Peano axioms,
as you found out doing this question.
If you want to see how to answer this question, and in general show the von Neumann
construction really works, I suggest you wait till you have time (probably in Summer) and
look it up in a book about the foundations of mathematics.

You might also like