# Phil.

**427/627: Lecture 2 (January 31, 2013)
**

Chapter 1: Enumerability

Goal of Chapter 1:

(i) To understand the concept of enumerability.

(ii) To understand the notion of partial function.

enumerable = countable

Idea: An enumerable set is one whose elements can be arranged in a sinlge list. Every memeber

appears sooner or later, maybe more than once.

(Note: Every ﬁnite set is enumerable.)

Deﬁnition: A denumerable set is an inﬁnte enumerable set.

Cardinal numbers

What do we mean by a set having 5 members? We will try to anwer this obvious question

by using a concept of function, since that will help our further discussions. A set A having ﬁve

members means that there is a one-to-one function from set ¦0, . . . , 4¦ onto set A. So, if a set A

has n memebrs, then there is a bijection f : ¦0, . . . , n − 1¦ −→ A.

We can extend this modeling (of the notion of size) from ﬁnite sets to inﬁnite sets, answering

the following questions: How can we talk about the size of an inﬁnte set? Is there one kind of

inﬁnity?

Deﬁnitions: Let A and B be sets.

1. We say that A and B have the same cardinality (write |A| = |B|) if and only if there is a

bijection f : A −→ B.

2. We write |A| ≤ |B| if there is one-to-one function f : A −→ B.

3. We say that |A| < |B| if |A| ≤ |B| and |A| , = |B|.

(Note: Having the same cardinality is an equivalence relation.)

Deﬁnition: A set is ﬁnite if it has the same cardinality as set ¦0, . . . , n¦ for some natural number

n ∈ A. Otherwise, the set is inﬁnite.

By the way, set A is the smallest(?) inﬁnite set, we call its cardinality ℵ

0

.

Deﬁnition:

A set A is countable iﬀ |A| ≤ ℵ

0

.

A set A is denumerable (or inﬁnitely enumerable or countably inﬁnite) iﬀ |A| = ℵ

0

.

A set A is uncountable iﬀ |A| > ℵ

0

.

(Examples) The following sets are denumerable sets:

Set of negative integers

Set of positive even numbers

Set of positive rational numbers

(Note: The laws of arithmetic for inﬁnite sets are diﬀerent than they are for ﬁnite sets.)

After understing the notion of partial function, we will come back to enumerability again.

1

Partial Function

We deﬁned a function to be a set of ordered pairs with the following characteristic:

for any x, there is one and only y such that ¸x, y¸ is in the set. That is,

f : A −→ B iﬀ Dom(f) = A and Rng(f) ⊆ B.

How about +(Mary, 3)?

Deﬁnition: f is deﬁned at x (write f(x)↓) iﬀ x ∈ Dom(f). [i.e. ∃yf(x) = y]

f is undeﬁned at x (write f(x)↑) iﬀ x ,∈ Dom(f).

Deﬁnition: f is a total function on A iﬀ f is a function and Dom(f) = A.

f is a partial function on A iﬀ f is a function, Dom(f) ⊆ A.

Important: Partial functions aren’t really a special type of function, but oﬀer a new and useful

way of considering old functions. Hence, it will be useful to have a way of talking about functions

without saying whether they are total or partial.

Deﬁnition: f; A −→ B(f semi-maps A into B) iﬀ Dom(f) ⊆ A and Rng(f) ⊆ B.

[I.e. ∃A

0

⊆ A such that f : A

0

−→ B]

Next, let’s extend the notions of 1-1 and onto mappings to semi-mappings:

Deﬁnition: f; A −→

1−1

B iﬀ ∃A

0

⊆ A such that f : A

0

−→

1−1

B

f; A −→

onto

B iﬀ ∃A

0

⊆ A such that f : A

0

−→

onto

B

Henceforth, the term ‘function’ will mean total or partial. That is, when we talk about a function

from A to A, we will mean a semi-mapping.

With a notion of partial functions, let’s continue our discussion of enumerability.

Redundancy in a list is okay, since we can always get rid of repetition.

How about gaps?

For example, 0, - , 1, -, 2, -, 3, . . . (Call it a gappy list.)

How can we show that this is enumerable list? That is, how can we make a function from the

set of natural numbers to this gappy list? This is where we will use a partial function.

f(0) = 0, f(1)↑, f(2) = 1, f(3)↑, . . .

Also,

f(n) = n/2 if n is even

↑ otherwise

The following function is the same as we did in the above, but let’s remember that f doest not

have to be total on A.

Deﬁnition: f enumerates A iﬀ f; A −→

1−1

onto

A

Deﬁnition: A is enumerable iﬀ ∃f such that f enumerates A.

[p.7: A is enumerable iﬀ it is the range of some function of natural numbers.]

Let’s compare this deﬁnition with the above deﬁnition ‘countable.’ That is,

A set A is countable iﬀ |A| ≤ ℵ

0

.

That is, A is countable iﬀ ∃X ⊆ A|A| = |X|.

Then,

|A| = |X| iﬀ ∃f f : A −→

1−1

onto

X

iﬀ ∃f f : X −→

1−1

onto

A

iﬀ ∃f f ; N −→

1−1

onto

A

2

Chapter 2: Diagonalization

Goal of Chapter 2:

(i) To see that some sets are not enumerable, and

(ii) to understand the technique involved in this kind of a proof.

Here is a time-honored theorem:

Cantor’s theorem: For any set A, |A| < |℘(A)|.

Proof : We need to show (i) |A| ≤ |℘(A)| and (ii) |A| , = |℘(A)|.

Show (i): To ﬁnd a one-to-one function f : A −→ ℘(A). Let f(x) = ¦x¦. This function is 1-1.

Show (ii): To show that there is no one-to-one onto function from A to ℘(A). Suppose that

there is a bijection f : A −→ ℘(A). [Show that there is a contradiction.] Since f is an onto

function, Rng(f) = ℘(A).

Let B = ¦x [ x ∈ A and x ,∈ f(x)¦. By the deﬁnition of this set, for every x ∈ B, x ∈ A. That

is, B ⊆ A. Accordingly, B ∈ ℘(A). Since Rng(f) = ℘(A), B ∈ Rng(f). Then, by the deﬁnition

of Rng(f), there is a such that f(a) = B where a ∈ A.

Is a ∈ B? If so, a ,∈ f(a) by the deﬁnition of B. That is, a ,∈ B, since f(a) = B. Contradiction.

Is a ,∈ B? If so, a ,∈ f(a) since f(a) = B. Then, by the deﬁnition of B, a ∈ B. Contradiction.

2

We can illustrate the same point in a slighly diﬀerent(?) way, which the book does.

The set of all sets of natural numbers (i.e. the set of the subsets of A) is not enumerable. It is

too big(!). In the proof for this proposition, we will learn a very useful technique: For a given any

list L of sets of natural numbers, we will construct a set D

∗

(L) of natural numbers which does

not show up in the list L. What if we add D

∗

(L) to the list? Then, for this new list L

, we can

come up with a set D

∗

(L

) which does not occur in the list L

.

Important Method: Diagnolization (How to construct D

∗

(L), given the list L.)

Given the list L of the sets of natural numbers,

S

1

, S

2

, S

3

, . . .

(Note: An inﬁnite set is included in the list as well.)

We will show this list is always missing an element, that is, we cannot enumerate them. We

suggest the following element as the candidate:

For each n, n ∈ D

∗

(L) iﬀ n ,∈ S

n

. (Note: D

∗

(L) is inﬁnite.)

(Example) ¦1¦, c, ¦1,2¦, c − ¦2¦, ¦1,3¦, . . .

S

1

, S

2

, S

3

, S

4

, S

5

, . . .

Then, D

∗

(L) = ¦3, 5, . . . ¦ (since 1 ∈ S

1

, 2 ∈ S

2

, 3 ,∈ S

3

, 4 ∈ S

4

, 5 ,∈ S

5

, . . . )

Show that D

∗

(L) is not in the list L.

Proof : Suppose by reductio that D

∗

(L) appears on the list. Say, S

m

= D

∗

(L).

Then, by the deﬁnition of D

∗

(L), we know that m ∈ D

∗

(L) iﬀ m ,∈ S

m

.

Since S

m

and D

∗

(L) are the same set, that m ∈ D

∗

(L) iﬀ m ,∈ D

∗

(L).

This is a plain contradiction!

Let’s look at this important diagonalization method from a slightly diﬀerent point of view:

3

Characteristic function

For a given set A, the characteristic function (

A

is a function that outputs 1 if its argument is in

the set A and 0 if it is not. That is,

(

A

(x) = 1 if x ∈ A

0 if x ,∈ A.

We can set up the characteristic function (

Sn

for each set of natural numbers, S

n

, in the following

way:

(

Sn

(x) = 1 if x ∈ S

n

0 if x ,∈ S

n

.

(Note: In B & J, the characteristic functions s

n

is the same as (

Sn

in the above.)

Then, we can draw the following array for the given list L: (Figure 2-1 on p.12)

1 2 3 4 . . .

(

S

1

(

S

1

(1) (

S

1

(2) (

S

1

(3) (

S

1

(4) . . .

(

S

2

(

S

2

(1) (

S

2

(2) (

S

2

(3) (

S

2

(4) . . .

(

S

3

(

S

3

(1) (

S

3

(2) (

S

3

(3) (

S

3

(4) . . .

(

S

4

(

S

4

(1) (

S

4

(2) (

S

4

(3) (

S

4

(4) . . .

.

.

.

Let’s go back to our example to see how the array looks like:

(Example) ¦1¦, c, ¦1,2¦, c − ¦2¦, ¦1,3¦, . . .

S

1

, S

2

, S

3

, S

4

, S

5

, . . .

1 2 3 4 . . .

(

S

1

1 0 0 0 . . .

(

S

2

0 1 0 1 . . .

(

S

3

1 1 0 0 . . .

(

S

4

0 0 0 1 . . .

(

S

5

1 0 1 0 . . .

.

.

.

(Note: From this array, we can tell what the original list is. For example, by looking at the

ﬁrst row, we know that S

1

= ¦1¦, . . . , by the ﬁfth row, S

5

= ¦1, 3¦. etc.

Let’s go back to our array, and draw a diagonal sequence like the following:

d = (

S

1

(1), (

S

2

(2), (

S

3

(3), (

S

4

(4), . . .

This diagonal sequence might be on the list. So, we need one more trick to make sure that we

create a sequence which does not appear on the list. We know that each element of sequence d is

either 1 or 0. So, we reverse each element of d by changing 1 to 0 and 0 to 1. So,

d

= 1 − (

S

1

(1), 1 − (

S

2

(2), 1 − (

S

3

(3), 1 − (

S

4

(4), . . .

4

Back to our example:

d = 1, 1, 0, 1, . . .

d

= 0, 0, 1, 0, . . .

We claim that d

**does not appear on the list L.
**

Proof : Suppose (by reductio) it does. Let’s say that d

**is the mth row in the array. Mth row
**

looks like this:

(

Sm

(

Sm

(1) (

Sm

(2) (

Sm

(3) (

Sm

(4) . . . . . . (

Sm

(m) . . . . . .

Since d

= 1−(

S

1

(1), 1−(

S

2

(2), 1−(

S

3

(3), 1−(

S

4

(4), . . . , 1−(

Sm

(m), . . . , we get the following

equations:

(

Sm

(1) = 1 − (

S

1

(1),

(

Sm

(2) = 1 − (

S

2

(2),

(

Sm

(3) = 1 − (

S

3

(3),

(

Sm

(4) = 1 − (

S

4

(4),

.

.

.

(

Sm

(m) = 1 − (

Sm

(m),

.

.

.

That is,

1 2 3 4 . . . m . . .

(

S

1

1 0 0 0 . . .

(

S

2

0 1 0 1 . . .

(

S

3

1 1 0 0 . . .

(

S

4

0 0 0 1 . . .

(

S

5

1 0 1 0 . . .

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

(

Sm

1 − (

S1

(1) 1 − (

S

2

(2) 1 − (

S

3

(3) 1 − (

S

4

(4) . . . 1 − (

Sm

(m)

.

.

.

But, (

Sm

(m) = 1 − (

Sm

(m) is a plain contradiction, which leads to 0=1! So, sequence d

cannot

appear on the list. 2

5