1 00:00:00,000 --> 00:00:03,300 [SOUND] Fundamentally, limits are promises. 2 00:00:03,300 --> 00:00:08,564 When I tell you the limit of F of X equals L, as X approaches A; I'm 3 00:00:08,564 --> 00:00:13,593 promising you something. I'm promising you that I can get F of X 4 00:00:13,593 --> 00:00:18,308 as close to L as you like as long as X is close enough to A. 5 00:00:18,308 --> 00:00:24,280 Thinking of limits as promises helps us to understand statements like these. 6 00:00:24,280 --> 00:00:28,768 Let's suppose that you tell me that you know the limit of f of x as x approaches 7 00:00:28,768 --> 00:00:31,484 a is equal to something, maybe this is equal to l. 8 00:00:31,484 --> 00:00:34,920 And maybe the limit of g of x as x approaches a is equal to m. 9 00:00:34,920 --> 00:00:40,120 What that really is, is a promise that you can make F of X as close to L as I 10 00:00:40,120 --> 00:00:43,294 like. And it's a promise that you can make G of 11 00:00:43,294 --> 00:00:48,495 X as close to M as I like, as long as I'm willing to make X close enough to A. 12 00:00:48,495 --> 00:00:53,830 But if you can promise me that you can make F of X close to L, and G of X close 13 00:00:53,830 --> 00:00:57,005 to M. Then I can turn back and promise you that 14 00:00:57,005 --> 00:01:00,720 F of X plus G of X is as close to L plus M as you like. 15 00:01:00,720 --> 00:01:03,442 Alright? If I want to make this close to 16 00:01:03,442 --> 00:01:08,547 something, I just ask you to make F of X close enough to L, and G of X close 17 00:01:08,547 --> 00:01:11,882 enough to M. So that F of X plus G of X, is as you 18 00:01:11,882 --> 00:01:15,540 like to L plus M. In other words, the limit of a sum is the 19 00:01:15,540 --> 00:01:18,492 sum of the limits. And the limit of a difference is the 20 00:01:18,492 --> 00:01:22,088 difference of the limits. And the limit of a product is the product 21 00:01:22,088 --> 00:01:25,258 of the limits. And what about quotient. 22 00:01:25,258 --> 00:01:28,481 And something similar is true for division. 23 00:01:28,481 --> 00:01:31,929 If the limit of F of X as X approaches A is L. 24 00:01:31,929 --> 00:01:38,074 And the limit of G of X as X approaches A is M, which isn't zero, then the limit of 25 00:01:38,074 --> 00:01:41,747 F of X over G of X, as X approaches A is L over M. 26 00:01:41,747 --> 00:01:47,743 In other words, the limit of the quotient is the quotient of the limits, provided 27 00:01:47,743 --> 00:01:52,540 those limits exist and, the limit of the denominator is non-zero. 28 00:01:52,540 --> 00:01:57,249 Let us do something with our new found knowledge about limits of quotients. 29 00:01:57,249 --> 00:02:01,645 Here's a limit problem: I'm going to limit x-squared over x plus one as x 30 00:02:01,645 --> 00:02:05,538 approaches two, alright? I'll promise you that x-squared over x 31 00:02:05,538 --> 00:02:09,620 plus one is close to something whenever x is close enough to two. 32 00:02:09,620 --> 00:02:14,848 This is the limit of a quotient, and the limit of the quotient's the quotient of 33 00:02:14,848 --> 00:02:19,619 the limits, provided the limit of the denominator is not zero, and in this 34 00:02:19,619 --> 00:02:22,887 case, it's not. So the limit of the quotient is the 35 00:02:22,887 --> 00:02:26,482 quotient of the limits. [SOUND] Here's the limit of the 36 00:02:26,482 --> 00:02:29,750 numerator. The limit is X approaches two of X plus 37 00:02:29,750 --> 00:02:32,430 one. This is the limit of the denominator. 38 00:02:32,430 --> 00:02:36,804 Now this is the limit of x squared. X squared is X times X. 39 00:02:36,804 --> 00:02:41,577 This is a limit of a product. And the limit of a products the product 40 00:02:41,577 --> 00:02:46,917 to the limits so I can replace the limit of the numerator with a limit of X as X 41 00:02:46,917 --> 00:02:50,609 approaches two times the limit of X as X approaches two. 42 00:02:50,609 --> 00:02:54,895 because this is the limit of X times X and here's the product limits. 43 00:02:54,895 --> 00:02:58,323 Limit of X times the limit of X as X approaches two. 44 00:02:58,323 --> 00:03:03,664 The denominator here is the limit of X plus one as X approaches two but that's a 45 00:03:03,664 --> 00:03:07,950 limit of a sum and the limit of the sum is the sum of the limits. 46 00:03:07,950 --> 00:03:16,076 So the limit of X plus one [SOUND] is the limit of X plus the limit of one, as X 47 00:03:16,076 --> 00:03:19,090 approaches two. Lets keep going. 48 00:03:19,090 --> 00:03:23,129 So I've got the limit of X times the limit of X over the limit of X plus the 49 00:03:23,129 --> 00:03:25,910 limit of one. And all of these limits are being taken, 50 00:03:25,910 --> 00:03:28,524 as X approaches two. What's the limit of X? 51 00:03:28,524 --> 00:03:34,357 That's asking what can you guarantee X is close to if you're willing to have X be 52 00:03:34,357 --> 00:03:38,482 close enough to two? Two is the limit of X as X goes to two. 53 00:03:38,482 --> 00:03:43,957 So limit of X as X goes to two is two. The limit of X as X goes to two is doubt 54 00:03:43,957 --> 00:03:49,571 is for multiplying, divided by limit of X as X approaches two plus with the limit 55 00:03:49,571 --> 00:03:54,214 of one as X approaches two. This is asking, what can I guarantee one 56 00:03:54,214 --> 00:03:58,719 is close to two if I am willing to have X be close enough to two. 57 00:03:58,719 --> 00:04:04,263 Well, one is already close to one, right. The limit of a constant is that constant. 58 00:04:04,263 --> 00:04:07,159 So this is just one. Two times two is four. 59 00:04:07,159 --> 00:04:11,344 Two plus one is three. And so the limit of this expression is 60 00:04:11,344 --> 00:04:15,049 four-thirds. At this point you are asking yourself why 61 00:04:15,049 --> 00:04:20,126 is the rule for limits of quotients different than limits of the products. 62 00:04:20,126 --> 00:04:23,419 A limit of a product is a product of the limits. 63 00:04:23,419 --> 00:04:27,110 Laws of limits exist. Why do I have to worry about the limit of 64 00:04:27,110 --> 00:04:31,300 a nominator being non-zero, when I'm taking the limit of a quotient? 65 00:04:31,300 --> 00:04:34,544 Most basically the problem is that you can't divide by zero. 66 00:04:34,544 --> 00:04:38,925 You can't go around telling people that the limit of a quotient is a quotient of 67 00:04:38,925 --> 00:04:43,089 the limits because the limit of the denominator might be zero and then you'd 68 00:04:43,089 --> 00:04:46,117 be telling people to divide by zero, which they can't do. 69 00:04:46,117 --> 00:04:49,878 You can't divide by zero. But you can think about it even a little 70 00:04:49,878 --> 00:04:52,670 more subtlety. You know, let's kind of unpack this a 71 00:04:52,670 --> 00:04:54,914 bit. Here's an example to think about: the 72 00:04:54,914 --> 00:04:57,651 limit of x over x minus three as x approaches six. 73 00:04:57,651 --> 00:05:01,263 This is no problem, alright? The numerator is a number close to six, 74 00:05:01,263 --> 00:05:05,533 it's how we're thinking about it, and the denominator is a number close to six 75 00:05:05,533 --> 00:05:08,341 minus three. A number close to six minus three, that 76 00:05:08,341 --> 00:05:11,044 means the denominator is a number close to three. 77 00:05:11,044 --> 00:05:14,961 Now, we've got a number close to six divided by a number close to three. 78 00:05:14,961 --> 00:05:18,933 Well, that's a number close to two. And, indeed, I mean, this limit is equal 79 00:05:18,933 --> 00:05:21,580 to two. I can make this quotient as close to two 80 00:05:21,580 --> 00:05:24,449 as I like. Because I can make the numerator as close 81 00:05:24,449 --> 00:05:28,697 to six as I need, the denominator is as close to three as I need, to guarantee 82 00:05:28,697 --> 00:05:31,290 that this ratio is as close to two as you like. 83 00:05:31,290 --> 00:05:36,202 So that limit is two. But what if instead of asking about the 84 00:05:36,202 --> 00:05:42,000 limit as x approaches six, I'd ask about the limit as x approaches three. 85 00:05:42,000 --> 00:05:45,772 Well, then what would I know? Then I'd know that the numerator was a 86 00:05:45,772 --> 00:05:49,714 number close to three, and the denominator was a number close to three 87 00:05:49,714 --> 00:05:52,657 minus three. The denominators aren't close to zero. 88 00:05:52,657 --> 00:05:57,071 The limited denominator is zero. That's exactly the scenario that the rule 89 00:05:57,071 --> 00:06:00,949 for taking limits of quotients is forbidding us from considering. 90 00:06:00,949 --> 00:06:05,424 We're not allowed to use the rule for limits of quotients here because the 91 00:06:05,424 --> 00:06:08,705 limited denominator is zero. But what really goes wrong? 92 00:06:08,705 --> 00:06:12,822 I mean, yeah, I can't divide by zero. Fine, I'm not going to divide by zero, 93 00:06:12,822 --> 00:06:15,387 I'm just dividing by numbers close to zero. 94 00:06:15,387 --> 00:06:18,728 But what happens when I divide by numbers close to zero? 95 00:06:18,728 --> 00:06:22,188 A number close to three divided by a number close to zero. 96 00:06:22,188 --> 00:06:26,671 Is that close to anything? If a number's close to zero it might be 97 00:06:26,671 --> 00:06:31,434 positive, and very small. Three divided by a small positive number 98 00:06:31,434 --> 00:06:36,380 is a huge positive number. What if the denominator were a number 99 00:06:36,380 --> 00:06:40,824 close to zero but negative? Very small, negative number close to 100 00:06:40,824 --> 00:06:43,857 zero. Three divided by a small but negative 101 00:06:43,857 --> 00:06:47,101 number? That would be a hugely negative number. 102 00:06:47,101 --> 00:06:50,558 Every negative. Well, this means this thing should be 103 00:06:50,558 --> 00:06:55,109 getting close to both a gigantic positive number and a very, very negative number. 104 00:06:55,109 --> 00:06:57,525 This thing isn't getting close to anything. 105 00:06:57,525 --> 00:07:00,503 Fundamentally that's the problem but not all is lost. 106 00:07:00,503 --> 00:07:05,054 Think about this slightly harder example: the limit of x-squared minus one over x 107 00:07:05,054 --> 00:07:08,261 minus one as x approaches one. It's a limit of a quotient. 108 00:07:08,261 --> 00:07:12,711 So your first temptation is to replace the limit of a quotient, by the quotient 109 00:07:12,711 --> 00:07:16,596 of the limits. But you can't replace it without quotient 110 00:07:16,596 --> 00:07:20,651 of limits, because the limit of the denominator is zero. 111 00:07:20,651 --> 00:07:24,632 The limit of X-1 as X approaches one is equal to zero. 112 00:07:24,632 --> 00:07:28,097 So it seems like our limit laws have failed us. 113 00:07:28,097 --> 00:07:31,783 And then we have got one more trick up our sleeve. 114 00:07:31,783 --> 00:07:35,838 Look at the numerator. X squared - one that factors as X+1 115 00:07:35,838 --> 00:07:38,271 times, X-1. First record that fact. 116 00:07:38,271 --> 00:07:43,580 And I am going to record the fact that the numerator factor is X+1 times X-1. 117 00:07:43,580 --> 00:07:48,980 Now I've still got a limit of a quotient and the limit of the denominator is still 118 00:07:48,980 --> 00:07:51,192 zero. So it seems like we're stuck. 119 00:07:51,192 --> 00:07:56,527 But now I've got a factor of X minus one in the numerator and a factor of X minus 120 00:07:56,527 --> 00:07:59,586 one in the denominator, and I can use that fact. 121 00:07:59,586 --> 00:08:02,904 What I want to do is imagine canceling these, right. 122 00:08:02,904 --> 00:08:08,044 I'd like to write that this is equal to the limit of just X plus one, as X goes 123 00:08:08,044 --> 00:08:11,452 to one. But note, these are not actually the same 124 00:08:11,452 --> 00:08:12,676 function. Alright? 125 00:08:12,676 --> 00:08:17,717 This thing up here is not defined at one, this thing is defined at one. 126 00:08:17,717 --> 00:08:22,974 And yet the limit doesn't care. The limit only depends upon values of the 127 00:08:22,974 --> 00:08:27,078 function near one. And near one, this and this are exactly 128 00:08:27,078 --> 00:08:30,607 the same. I'm not allowed to plug one into this, I 129 00:08:30,607 --> 00:08:35,280 am allowed to plug one into this. They're different functions. 130 00:08:35,280 --> 00:08:40,605 But those functions are equal if you're only considering values that are near and 131 00:08:40,605 --> 00:08:44,104 not equal to one. As a result, these limits are the same. 132 00:08:44,104 --> 00:08:49,083 This is another limit of a sum, and the limit of a sum is the sum of the limits, 133 00:08:49,083 --> 00:08:53,369 so this is the limit of X plus the limit of one as X approaches one. 134 00:08:53,369 --> 00:08:56,016 The limit of X as X approaches one is one. 135 00:08:56,016 --> 00:09:00,050 The limit of one, which is a constant, as X approaches one is one. 136 00:09:00,050 --> 00:09:02,130 This is one plus one. This is two. 137 00:09:02,130 --> 00:09:15,860 And so that limit is equal to two and our limit laws have again saved the day. 138 00:09:15,860 --> 00:09:17,269 [SOUND]