WEBVTT

00:00:00.000 --> 00:00:09.110
JOEL LEWIS: Hi, welcome
back to recitation.

00:00:09.110 --> 00:00:11.090
In lecture,
Professor Jerison was

00:00:11.090 --> 00:00:13.070
teaching about L'Hospital's
Rule, and one thing

00:00:13.070 --> 00:00:15.000
that he mentioned
several times was

00:00:15.000 --> 00:00:17.530
that when you apply
L'Hospital's Rule,

00:00:17.530 --> 00:00:20.780
it's really important that the
second limit exists in order

00:00:20.780 --> 00:00:22.430
for L'Hospital's
Rule to be true,

00:00:22.430 --> 00:00:24.721
in order for the first limit
to equal the second limit.

00:00:24.721 --> 00:00:26.620
So I have a-- but
he didn't give you

00:00:26.620 --> 00:00:31.710
any examples of
what can go wrong.

00:00:31.710 --> 00:00:35.780
So I have here a limit
for you to work on.

00:00:35.780 --> 00:00:39.490
So what I'd like you to do
is try L'Hospital's Rule,

00:00:39.490 --> 00:00:43.670
see what happens, and then try
to solve it a different way not

00:00:43.670 --> 00:00:45.240
using L'Hospital's Rule.

00:00:45.240 --> 00:00:47.697
So why don't you
pause the video,

00:00:47.697 --> 00:00:49.780
spend a couple minutes
working on that, come back,

00:00:49.780 --> 00:00:51.071
and we can work on it together.

00:00:59.300 --> 00:01:00.030
Welcome back.

00:01:00.030 --> 00:01:02.740
Hopefully you've have some
fun working on this limit.

00:01:02.740 --> 00:01:05.050
Let's go through it
and see what happens

00:01:05.050 --> 00:01:07.370
when we try to apply
L'Hospital's Rule to it.

00:01:07.370 --> 00:01:10.725
So the first thing to notice
is that as x goes to infinity,

00:01:10.725 --> 00:01:13.865
x plus cosine x-- well,
cosine x is small,

00:01:13.865 --> 00:01:16.750
and you know, between minus 1
and 1-- x is going to infinity.

00:01:16.750 --> 00:01:18.820
So the top is going to
infinity and the bottom

00:01:18.820 --> 00:01:19.780
is going to infinity.

00:01:19.780 --> 00:01:22.940
So this is an infinity over
infinity indeterminate form.

00:01:22.940 --> 00:01:24.670
So it is a context
in which we can

00:01:24.670 --> 00:01:27.530
try to apply L'Hospital's
Rule, so that's fine.

00:01:27.530 --> 00:01:30.220
So it's an
indeterminate ratio, we

00:01:30.220 --> 00:01:32.180
can look at applying
L'Hospital's Rule.

00:01:32.180 --> 00:01:33.870
So let's try to do that.

00:01:33.870 --> 00:01:38.730
So L'Hospital's Rule
says that the limit

00:01:38.730 --> 00:01:43.970
as x goes to infinity
of our expression x plus

00:01:43.970 --> 00:01:47.192
cosine x divided by
x is equal-- and I'm

00:01:47.192 --> 00:01:49.400
going to put a little question
mark here because what

00:01:49.400 --> 00:01:52.050
L'Hospital's Rule says is
that it's equal provided--

00:01:52.050 --> 00:01:54.050
it's equal to what I'm
going to write over here,

00:01:54.050 --> 00:01:56.000
provided that the
second limit exists.

00:01:56.000 --> 00:01:57.170
So what goes over here?

00:01:57.170 --> 00:02:01.886
Well it's the limit as x goes
to the same place, as x goes

00:02:01.886 --> 00:02:04.870
to infinity, of the
ratio of the derivative,

00:02:04.870 --> 00:02:07.290
so the derivative of the top
divided by the derivative

00:02:07.290 --> 00:02:08.220
of the bottom.

00:02:08.220 --> 00:02:13.080
So in this case
that's 1 minus sine

00:02:13.080 --> 00:02:16.350
x is the derivative
of the top and 1 is

00:02:16.350 --> 00:02:18.550
the derivative of the bottom.

00:02:18.550 --> 00:02:20.059
OK, now let's look
at this limit.

00:02:20.059 --> 00:02:20.600
What happens?

00:02:20.600 --> 00:02:23.430
Well, OK the over 1
part is irrelevant.

00:02:23.430 --> 00:02:26.060
The 1 minus part, OK, who cares?

00:02:26.060 --> 00:02:29.820
The only thing that's changing
in this limit is the sine x.

00:02:29.820 --> 00:02:33.120
So as x goes to infinity what
this does is it oscillates,

00:02:33.120 --> 00:02:34.270
just like sine x does.

00:02:34.270 --> 00:02:36.480
I mean, it's offset
and flipped upside down

00:02:36.480 --> 00:02:38.300
because of this 1 minus.

00:02:38.300 --> 00:02:40.870
So, in particular, you
know, sometimes it's near 1,

00:02:40.870 --> 00:02:43.220
sometimes it's near negative
1, sometimes it's near 0,

00:02:43.220 --> 00:02:46.760
and it oscillates back
and forth in between.

00:02:46.760 --> 00:02:48.920
So in particular,
because it's oscillating,

00:02:48.920 --> 00:02:50.370
it's not approaching any value.

00:02:50.370 --> 00:02:53.200
So this limit doesn't exist.

00:02:53.200 --> 00:02:54.655
It's not equal to
any real number.

00:02:54.655 --> 00:02:55.780
It's not equal to infinity.

00:02:55.780 --> 00:02:57.640
It's not equal to
minus infinity.

00:02:57.640 --> 00:02:58.714
It does not exist.

00:03:01.620 --> 00:03:05.210
OK, so the statement
of L'Hospital's theorem

00:03:05.210 --> 00:03:07.125
says that this equality
is really wrong.

00:03:07.125 --> 00:03:08.465
OK?

00:03:08.465 --> 00:03:10.090
What it means is that
L'Hospital's Rule

00:03:10.090 --> 00:03:11.230
tells you nothing.

00:03:11.230 --> 00:03:12.980
You don't learn anything here.

00:03:12.980 --> 00:03:16.220
So if we want to compute
this first limit,

00:03:16.220 --> 00:03:17.620
we can't use L'Hospital's Rule.

00:03:17.620 --> 00:03:21.820
We have to come up with some
better way to do it, OK.

00:03:21.820 --> 00:03:23.210
So that's what's going wrong.

00:03:23.210 --> 00:03:25.300
Now let's see about
doing this even

00:03:25.300 --> 00:03:27.460
though we don't have this tool.

00:03:27.460 --> 00:03:29.810
So if we try and
solve this limit

00:03:29.810 --> 00:03:32.520
without L'Hospital's Rule, so
we want to look at the limit

00:03:32.520 --> 00:03:36.040
as x goes to infinity of x
plus cosine x divided by x.

00:03:36.040 --> 00:03:39.070
Well, think about what's
important in this limit.

00:03:39.070 --> 00:03:44.230
As x is getting big, well
x is getting big and x

00:03:44.230 --> 00:03:46.460
is getting big and
what's cosine x doing?

00:03:46.460 --> 00:03:50.130
Well cosine x is behaving
like a constant but wigglier,

00:03:50.130 --> 00:03:51.620
right, we could say.

00:03:51.620 --> 00:03:54.240
So cosine x is always
between minus 1 and 1,

00:03:54.240 --> 00:03:56.200
but, you know, it's
oscillating, but it's

00:03:56.200 --> 00:04:01.480
in some bounded interval
there, whereas x and x are both

00:04:01.480 --> 00:04:03.760
going to infinity at
exactly the same speed,

00:04:03.760 --> 00:04:08.980
so this suggests a manipulation
that sort of allows

00:04:08.980 --> 00:04:11.400
us to quantify that explicitly.

00:04:11.400 --> 00:04:13.240
And so one thing
that we can do is

00:04:13.240 --> 00:04:16.870
we can divide top and
bottom of this limit by x.

00:04:16.870 --> 00:04:19.420
So if we do that, or
equivalently we can just

00:04:19.420 --> 00:04:22.600
divide this x into
the fraction above,

00:04:22.600 --> 00:04:24.050
split it into two fractions.

00:04:24.050 --> 00:04:27.520
So we have-- I'm just going
to re-write it one more

00:04:27.520 --> 00:04:33.490
time-- the limit as x goes
to infinity of x plus cosine

00:04:33.490 --> 00:04:39.900
x divided by x is
equal to the limit

00:04:39.900 --> 00:04:48.560
as x goes to infinity of
1 plus cosine x over x.

00:04:48.560 --> 00:04:50.110
All right, so what's
happening here?

00:04:50.110 --> 00:04:53.940
Well, all I've done is I've
divided that x in-- and now you

00:04:53.940 --> 00:04:56.590
see that this 1, that was
from the x over x, so that's

00:04:56.590 --> 00:04:58.020
a constant that's going to 1.

00:04:58.020 --> 00:05:00.440
And as x goes to
infinity, cosine x-- again

00:05:00.440 --> 00:05:02.996
we can think of it like
a constant but wigglier--

00:05:02.996 --> 00:05:04.370
x is going to
infinity, so cosine

00:05:04.370 --> 00:05:10.390
x over x is going to 0
as x goes to infinity.

00:05:10.390 --> 00:05:12.250
So I guess I wouldn't
call this plugging in,

00:05:12.250 --> 00:05:14.960
but it's, you know, sort
of just like plugging in.

00:05:14.960 --> 00:05:17.440
So this part is going to 0,
so we're left with just 1.

00:05:20.530 --> 00:05:23.870
OK, so we can evaluate
this limit fairly easily.

00:05:23.870 --> 00:05:25.440
It just took a couple steps.

00:05:25.440 --> 00:05:27.890
But we can't evaluate it
with L'Hospital's Rule.

00:05:27.890 --> 00:05:32.410
So this is why you have that
extra thing in the statement

00:05:32.410 --> 00:05:35.020
of the theorem that Professor
Jerison kept saying,

00:05:35.020 --> 00:05:37.130
and that I'll say
every time I use it,

00:05:37.130 --> 00:05:40.424
which is L'Hospital's Rule says
that the limit of the ratio

00:05:40.424 --> 00:05:42.590
is equal to the limit of
the ratio of the derivative

00:05:42.590 --> 00:05:44.920
provided the second
limit exists.

00:05:44.920 --> 00:05:47.750
OK, so here's an example where
the second limit didn't exist

00:05:47.750 --> 00:05:49.920
and so we can't say that
the two are actually

00:05:49.920 --> 00:05:52.980
equal to each other, and we need
some other tools to evaluate

00:05:52.980 --> 00:05:54.330
the limit in question.

00:05:54.330 --> 00:05:55.656
So I'll end there.