1 00:00:01,120 --> 00:00:02,980 PROFESSOR: So let's look at two examples 2 00:00:02,980 --> 00:00:04,820 of using the well-ordering principle. 3 00:00:04,820 --> 00:00:06,820 One of them is pretty obvious, and the other one 4 00:00:06,820 --> 00:00:11,414 is not hard but a little bit more interesting. 5 00:00:11,414 --> 00:00:13,830 So what we're going to prove is that every integer greater 6 00:00:13,830 --> 00:00:16,770 than one is a product of primes. 7 00:00:16,770 --> 00:00:20,800 So remember a prime is an integer greater than 1 8 00:00:20,800 --> 00:00:27,060 that is only divisible by itself and the number 1. 9 00:00:27,060 --> 00:00:31,400 It can't be expressed as the product of other numbers 10 00:00:31,400 --> 00:00:33,630 greater than 1. 11 00:00:33,630 --> 00:00:36,470 So the way we're going to prove this is by contradiction, 12 00:00:36,470 --> 00:00:38,320 and we're going to begin by assuming 13 00:00:38,320 --> 00:00:41,570 suppose that there were some numbers that 14 00:00:41,570 --> 00:00:43,820 were non-products of primes. 15 00:00:43,820 --> 00:00:45,010 OK. 16 00:00:45,010 --> 00:00:48,550 That is to say, the set of non-products is non-empty. 17 00:00:48,550 --> 00:00:51,960 So applying the least-- the well-ordering principle 18 00:00:51,960 --> 00:00:53,950 to this non-empty set of non-products, 19 00:00:53,950 --> 00:00:58,380 there's got to be a least one, so m is a number greater than 1 20 00:00:58,380 --> 00:01:01,080 that is not a product of primes. 21 00:01:01,080 --> 00:01:04,300 Now, by convention, if m itself was a prime, 22 00:01:04,300 --> 00:01:06,340 it's considered to be a product of one prime, 23 00:01:06,340 --> 00:01:09,105 so we know that m is not a prime. 24 00:01:14,871 --> 00:01:15,370 Now look. 25 00:01:15,370 --> 00:01:18,720 M is not a prime, or if it was a prime, 26 00:01:18,720 --> 00:01:20,770 it would be a product of just itself, 27 00:01:20,770 --> 00:01:24,460 so that means that it must be a product of two numbers, 28 00:01:24,460 --> 00:01:27,770 call them j and k, where j and k are greater than 1 29 00:01:27,770 --> 00:01:28,505 and less than m. 30 00:01:28,505 --> 00:01:30,380 That's what it means to be a non-prime-- it's 31 00:01:30,380 --> 00:01:32,460 a product of j and k. 32 00:01:32,460 --> 00:01:36,480 Well, j and k are less than m, so that 33 00:01:36,480 --> 00:01:40,160 means that they must be prime products, because they're 34 00:01:40,160 --> 00:01:43,090 less than m and greater than 1, and m 35 00:01:43,090 --> 00:01:46,040 is the smallest such number that's not a product of primes. 36 00:01:46,040 --> 00:01:49,550 So we can assume that j is equal to some product of prime say, 37 00:01:49,550 --> 00:01:54,030 p1 through p94, and k is some other product of primes, 38 00:01:54,030 --> 00:01:58,620 q1 through q13, so you can see where this is going. 39 00:01:58,620 --> 00:02:02,170 Now what we have is that m, which is jk, 40 00:02:02,170 --> 00:02:04,760 is simply the product of those p's 41 00:02:04,760 --> 00:02:07,820 followed by the product of those q's and is, 42 00:02:07,820 --> 00:02:11,070 in fact, a prime product which is a contradiction. 43 00:02:11,070 --> 00:02:13,970 So what did we assume that led to the contradiction? 44 00:02:13,970 --> 00:02:16,240 We assumed that there were some counter-examples, 45 00:02:16,240 --> 00:02:19,710 and there must not be any, and no counter-examples 46 00:02:19,710 --> 00:02:23,210 means that, in fact, every single integer greater than 1 47 00:02:23,210 --> 00:02:26,487 is indeed a product of primes as [AUDIO OUT]. 48 00:02:26,487 --> 00:02:28,570 Let's start looking at a slightly more interesting 49 00:02:28,570 --> 00:02:30,850 example using the well-ordered principle 50 00:02:30,850 --> 00:02:33,280 to reasoning about postage. 51 00:02:33,280 --> 00:02:35,310 So suppose that we have a bunch of $0.05 52 00:02:35,310 --> 00:02:38,630 stamps and $0.03 stamps, and what 53 00:02:38,630 --> 00:02:41,260 I want to analyze is what amounts of postage 54 00:02:41,260 --> 00:02:44,380 can you make out of $0.05 stamps and $0.03 stamps? 55 00:02:44,380 --> 00:02:46,750 So I'm going to introduce a technical definition 56 00:02:46,750 --> 00:02:48,630 for convenience. 57 00:02:48,630 --> 00:02:52,480 Let's say that a number n is postal. 58 00:02:52,480 --> 00:02:56,160 If I can make n plus $0.08 postage from $0.03 59 00:02:56,160 --> 00:03:00,550 and $0.05 stamps. 60 00:03:00,550 --> 00:03:02,760 So this is what I'm going to prove. 61 00:03:02,760 --> 00:03:04,850 I claim that every number is postal. 62 00:03:04,850 --> 00:03:08,200 In other words, I can make every amount of postage 63 00:03:08,200 --> 00:03:11,490 from $0.08 up. 64 00:03:11,490 --> 00:03:13,160 I'm going to prove this by applying 65 00:03:13,160 --> 00:03:14,730 the well-ordering principle, and as 66 00:03:14,730 --> 00:03:16,250 usual with well-ordering principles 67 00:03:16,250 --> 00:03:18,690 we'll begin by supposing that there was 68 00:03:18,690 --> 00:03:20,590 a number that wasn't postal. 69 00:03:20,590 --> 00:03:23,340 That would be a counter-example, so if there's 70 00:03:23,340 --> 00:03:26,180 any number that's not postal, then there's 71 00:03:26,180 --> 00:03:29,932 at least one m by the well-ordering principal, 72 00:03:29,932 --> 00:03:31,390 because the set of counter-examples 73 00:03:31,390 --> 00:03:34,180 is non-empty if some number is not postal, 74 00:03:34,180 --> 00:03:36,830 so there's at least one. 75 00:03:36,830 --> 00:03:38,550 So what we know, in other words, is 76 00:03:38,550 --> 00:03:41,690 that this least m that's not postal has the property. 77 00:03:41,690 --> 00:03:47,450 It's not postal, and any number less than it is postal. 78 00:03:47,450 --> 00:03:49,230 See what we can figure out about m. 79 00:03:49,230 --> 00:03:54,840 First of all, m is not 0-- 0 is postal, because 0 plus $0.08 80 00:03:54,840 --> 00:03:58,130 can be made with a $0.03 stamp and a $0.05 stamp. 81 00:03:58,130 --> 00:04:04,050 M is not 0, because m is supposed to be not postal. 82 00:04:04,050 --> 00:04:05,850 As a matter of fact, by the same reasoning, 83 00:04:05,850 --> 00:04:10,780 m is not 1 because you can make 1 plus $0.08 with three $0.03, 84 00:04:10,780 --> 00:04:15,630 and m is not 2, because you can make 2 plus $0.08-- $0.10-- 85 00:04:15,630 --> 00:04:17,800 using two $0.05. 86 00:04:17,800 --> 00:04:22,365 So we've just figured out that this least counter-example has 87 00:04:22,365 --> 00:04:23,740 to be greater than or equal to 3, 88 00:04:23,740 --> 00:04:27,300 because 0, 1, and 2 are not counter-examples. 89 00:04:27,300 --> 00:04:28,780 So we've got that m is greater than 90 00:04:28,780 --> 00:04:31,280 or equal to 3, the least non-postal number, 91 00:04:31,280 --> 00:04:33,271 so if I look at m minus 3, that means 92 00:04:33,271 --> 00:04:35,270 it's a number that's greater than or equal to 0, 93 00:04:35,270 --> 00:04:38,060 and it's less than m, so it's postal, 94 00:04:38,060 --> 00:04:40,810 because m is the least non-postal one. 95 00:04:40,810 --> 00:04:43,900 So, in other words, I can make-- out of $0.03 and $0.05 stamps, 96 00:04:43,900 --> 00:04:47,520 I can make m minus 3 plus $0.08 but, look, 97 00:04:47,520 --> 00:04:49,800 if I can make m minus 3 plus $0.08, 98 00:04:49,800 --> 00:04:53,190 then obviously m is postal also, because I just 99 00:04:53,190 --> 00:04:57,300 add $0.03 to that and minus 3 number, 100 00:04:57,300 --> 00:05:00,250 and I wind up with m plus $0.08, which 101 00:05:00,250 --> 00:05:02,850 says that m is postal and is a contradiction. 102 00:05:02,850 --> 00:05:06,380 So assuming that there was a least non-postal number, 103 00:05:06,380 --> 00:05:09,120 I reach a contradiction and therefore there 104 00:05:09,120 --> 00:05:11,360 is no non-postal number. 105 00:05:11,360 --> 00:05:14,810 Every number is postal-- 0 plus 8 is postal, 106 00:05:14,810 --> 00:05:17,880 1 plus 8 is postal, 2 plus 8 is postal. 107 00:05:17,880 --> 00:05:20,020 Every number greater than or equal to $0.08 108 00:05:20,020 --> 00:05:22,999 can be made out of $0.03 and $0.05 stamps.