WEBVTT
00:00:00.500 --> 00:00:03.800
For those of you who are
curious, we will go through an
00:00:03.800 --> 00:00:07.420
argument that establishes that
the set of real numbers is an
00:00:07.420 --> 00:00:08.830
uncountable set.
00:00:08.830 --> 00:00:11.060
It's a famous argument
known as Cantor's
00:00:11.060 --> 00:00:13.570
diagonalization argument.
00:00:13.570 --> 00:00:16.480
Actually, instead of looking
at the set of all real
00:00:16.480 --> 00:00:22.190
numbers, we will first look at
the set of all numbers, x,
00:00:22.190 --> 00:00:26.110
that belong to the open
unit interval--
00:00:26.110 --> 00:00:28.880
so numbers between 0 and 1--
00:00:28.880 --> 00:00:34.230
and such that their decimal
expansion involves
00:00:34.230 --> 00:00:36.335
only threes and fours.
00:00:40.980 --> 00:00:45.305
Now, the choice of three and
four is somewhat arbitrary.
00:00:45.305 --> 00:00:46.840
It doesn't matter.
00:00:46.840 --> 00:00:49.700
What really matters is
that we do not have
00:00:49.700 --> 00:00:52.280
long strings of nines.
00:00:52.280 --> 00:00:57.300
So suppose that this
set was countable.
00:00:57.300 --> 00:01:04.440
If the set was countable, then
that set could be written as
00:01:04.440 --> 00:01:11.720
equal to a set of this form,
x1, x2, x3 and so on, where
00:01:11.720 --> 00:01:18.280
each one of these is a real
number inside that set.
00:01:18.280 --> 00:01:20.860
Now, suppose that this
is the case.
00:01:20.860 --> 00:01:23.690
Let us take those numbers
and write them
00:01:23.690 --> 00:01:26.620
down in decimal notation.
00:01:26.620 --> 00:01:31.760
For example, one number
could be this one, and
00:01:31.760 --> 00:01:32.920
it continues forever.
00:01:32.920 --> 00:01:35.330
Since we're talking about real
numbers, their decimal
00:01:35.330 --> 00:01:38.210
expansion will go on forever.
00:01:38.210 --> 00:01:43.979
Suppose that the second number
is of this kind, and it has
00:01:43.979 --> 00:01:45.495
its own decimal expansion.
00:01:48.150 --> 00:01:53.430
Suppose that the third number
is, again, with some decimal
00:01:53.430 --> 00:01:55.690
expansion and so on.
00:01:55.690 --> 00:02:00.790
So we have assumed that our
set is countable and
00:02:00.790 --> 00:02:03.930
therefore, the set is equal
to that sequence.
00:02:03.930 --> 00:02:09.470
So this sequence exhausts all
the numbers in that set.
00:02:09.470 --> 00:02:12.270
Can it do that?
00:02:12.270 --> 00:02:16.220
Let's construct a new number
in the following fashion.
00:02:16.220 --> 00:02:19.579
The new number looks
at this digit and
00:02:19.579 --> 00:02:20.840
does something different.
00:02:20.840 --> 00:02:24.010
Looks at this digit, the second
digit of the second
00:02:24.010 --> 00:02:26.420
number, and does something
different.
00:02:26.420 --> 00:02:29.840
Looks at the third digit of
the third number and does
00:02:29.840 --> 00:02:30.980
something different.
00:02:30.980 --> 00:02:33.200
And we continue this way.
00:02:33.200 --> 00:02:36.829
This number that we have
constructed here is different
00:02:36.829 --> 00:02:38.120
from the first number.
00:02:38.120 --> 00:02:39.820
They differ in the
first digit.
00:02:39.820 --> 00:02:41.520
It's different from
the second number.
00:02:41.520 --> 00:02:43.470
They differ in the
second digit.
00:02:43.470 --> 00:02:46.370
It's different from the third
number because it's different
00:02:46.370 --> 00:02:48.780
in the third digit and so on.
00:02:48.780 --> 00:02:55.900
So this is a number, and this
number is different
00:02:55.900 --> 00:03:01.310
from xi for all i.
00:03:01.310 --> 00:03:07.510
So we have an element of this
set which does not belong to
00:03:07.510 --> 00:03:08.800
this sequence.
00:03:08.800 --> 00:03:13.970
Therefore, it cannot be true
that this set is equal to the
00:03:13.970 --> 00:03:16.980
set formed by that sequence.
00:03:16.980 --> 00:03:21.610
And so this is a contradiction
to the initial assumption that
00:03:21.610 --> 00:03:25.100
this set could be written
in this form, and this
00:03:25.100 --> 00:03:29.560
contradiction establishes that
since this is not possible,
00:03:29.560 --> 00:03:34.520
that the set that we have here
is an uncountable set.
00:03:34.520 --> 00:03:38.560
Now, this set is a subset of
the set of real numbers.
00:03:38.560 --> 00:03:42.030
Since this one is uncountable,
it is not hard to show that
00:03:42.030 --> 00:03:46.060
the set of real numbers, which
is a bigger set, will also be
00:03:46.060 --> 00:03:47.940
uncountable.
00:03:47.940 --> 00:03:51.440
And so this is this particular
famous argument.
00:03:51.440 --> 00:03:54.570
We will not need it or make any
arguments of this type in
00:03:54.570 --> 00:03:58.340
this class, but it's so
beautiful that it's worth for
00:03:58.340 --> 00:04:00.910
everyone to see it once
in their lifetime.