1 00:00:01,030 --> 00:00:04,840 PROFESSOR: We spoke last time of the existence 2 00:00:04,840 --> 00:00:06,430 of symmetric states. 3 00:00:11,260 --> 00:00:16,530 And for that we were referring to states 4 00:00:16,530 --> 00:00:25,890 psi S that belonged to the M particle Hilbert space. 5 00:00:25,890 --> 00:00:31,585 And V is the vector space that applies to the states of the M 6 00:00:31,585 --> 00:00:32,746 particles. 7 00:00:32,746 --> 00:00:36,180 And we constructed some states. 8 00:00:36,180 --> 00:00:38,600 So within the construction, we postulated, 9 00:00:38,600 --> 00:00:42,200 there will be some states that [AUDIO OUT] 10 00:00:42,200 --> 00:00:48,390 that meant that any permutation of [AUDIO OUT] state 11 00:00:48,390 --> 00:00:52,970 will be for the symmetric state or all alpha. 12 00:00:52,970 --> 00:00:57,590 So that meant the state was symmetric. 13 00:00:57,590 --> 00:01:01,210 Whatever you apply to it, any permutation, 14 00:01:01,210 --> 00:01:04,050 the state would be invariant. 15 00:01:04,050 --> 00:01:09,290 So these are special states, if they exist, in VN. 16 00:01:09,290 --> 00:01:12,445 Then there would be anti-symmetric states. 17 00:01:18,090 --> 00:01:22,960 And we concluded that an anti-symmetric state-- 18 00:01:22,960 --> 00:01:26,520 with an A for anti-symmetric-- 19 00:01:26,520 --> 00:01:32,500 also stayed in this tensor product. 20 00:01:32,500 --> 00:01:34,930 That state would react differently 21 00:01:34,930 --> 00:01:37,215 to the permutation operators. 22 00:01:37,215 --> 00:01:43,130 It would change up to a sine epsilon alpha times 23 00:01:43,130 --> 00:01:50,740 psi A. It would be an eigen state of all those permutation 24 00:01:50,740 --> 00:01:52,630 operators. 25 00:01:52,630 --> 00:01:57,140 But with eigenvalue epsilon and alpha 26 00:01:57,140 --> 00:01:59,920 where epsilon alpha was equal to plus 27 00:01:59,920 --> 00:02:05,710 1 if P alpha is an even permutation. 28 00:02:09,150 --> 00:02:15,540 Or minus 1 if P alpha is an odd permutation. 29 00:02:20,590 --> 00:02:24,540 So whether this permutation is even or odd, 30 00:02:24,540 --> 00:02:28,830 we also discussed depends on whether it's 31 00:02:28,830 --> 00:02:32,850 built with an even or odd number of transpositions. 32 00:02:32,850 --> 00:02:36,600 With transpositions being permutations in which 33 00:02:36,600 --> 00:02:39,870 one state is flipped for another state. 34 00:02:39,870 --> 00:02:42,610 Within the end states, you pick two, 35 00:02:42,610 --> 00:02:45,780 and these two are flipped, that's a transposition. 36 00:02:45,780 --> 00:02:50,250 All permutations can be built through transpositions-- 37 00:02:50,250 --> 00:02:52,410 with transpositions-- and therefore you 38 00:02:52,410 --> 00:02:56,250 can tell from a permutation whether it's 39 00:02:56,250 --> 00:03:00,060 an even or an odd one depending of whether it's 40 00:03:00,060 --> 00:03:03,360 built with even number of transpositions or odd number 41 00:03:03,360 --> 00:03:04,680 of transpositions. 42 00:03:04,680 --> 00:03:07,230 Now, the number of transpositions 43 00:03:07,230 --> 00:03:12,330 you need to build the permutation is not fixed. 44 00:03:12,330 --> 00:03:16,460 If we say it's even, means it's even mod 2. 45 00:03:16,460 --> 00:03:22,230 So you might have a permutation is built with 2 transpositions. 46 00:03:22,230 --> 00:03:24,580 And also somebody else can write it 47 00:03:24,580 --> 00:03:29,040 as 4 transpositions and 6 transpositions. 48 00:03:29,040 --> 00:03:32,230 It just doesn't matter. 49 00:03:32,230 --> 00:03:36,870 So a few facts that we learned about these things 50 00:03:36,870 --> 00:03:42,810 are that all the P alphas are unitary operators. 51 00:03:50,650 --> 00:03:59,050 And we also learned that all transpositions 52 00:03:59,050 --> 00:04:01,690 are Hermitian operators. 53 00:04:12,780 --> 00:04:18,490 Now, transpositions, of course, are permutations. 54 00:04:18,490 --> 00:04:22,570 So they're Hermitian, and they're unitary as well. 55 00:04:22,570 --> 00:04:32,970 And finally, we learned that the number of even permutations 56 00:04:32,970 --> 00:04:46,300 is equal to the number of odd permutations in any permutation 57 00:04:46,300 --> 00:04:47,665 group of n objects. 58 00:04:53,550 --> 00:04:58,470 So these were some of the facts we learned last time already. 59 00:04:58,470 --> 00:05:10,680 We have now that symmetric states form a subspace. 60 00:05:10,680 --> 00:05:12,690 If you have two symmetric states, 61 00:05:12,690 --> 00:05:15,400 you can multiply a state by a number, 62 00:05:15,400 --> 00:05:16,710 it will still be symmetric. 63 00:05:16,710 --> 00:05:20,350 If you add two symmetric states, will still be symmetric. 64 00:05:20,350 --> 00:05:25,440 So symmetric states form a subspace of the full vector 65 00:05:25,440 --> 00:05:28,600 space V tensor N. 66 00:05:28,600 --> 00:05:31,920 And anti-symmetric states also form 67 00:05:31,920 --> 00:05:37,180 a subspace of the vector space, V tensor N. 68 00:05:37,180 --> 00:05:40,180 So let's write these facts. 69 00:05:40,180 --> 00:05:54,460 So symmetric states form the subspace, 70 00:05:54,460 --> 00:06:07,600 and it's called sym N of V of V tensor N. Anti-symmetric 71 00:06:07,600 --> 00:06:30,540 states form the subspace anti N V of V tensor N. 72 00:06:30,540 --> 00:06:31,230 All right. 73 00:06:31,230 --> 00:06:34,440 So these are the states. 74 00:06:34,440 --> 00:06:37,950 But we have not learned how to build them, how to find them, 75 00:06:37,950 --> 00:06:41,980 and even more, what to do with them. 76 00:06:41,980 --> 00:06:47,130 So main thing is if this form subspaces, 77 00:06:47,130 --> 00:06:50,160 there should be a way to write the projector that 78 00:06:50,160 --> 00:06:53,880 takes from your big space down to the subspace. 79 00:06:53,880 --> 00:06:57,525 So here is the claim that we have. 80 00:07:00,660 --> 00:07:02,660 Here are two operators. 81 00:07:02,660 --> 00:07:12,390 We'll call this S, a symmetrizer And S 82 00:07:12,390 --> 00:07:18,540 will be billed as 1 over N factorial, the sum over alpha 83 00:07:18,540 --> 00:07:21,840 of all the permutations, P alpha. 84 00:07:21,840 --> 00:07:22,980 That's the definition. 85 00:07:30,880 --> 00:07:41,370 And then we'll have an anti-symmetrizer, A. This is 86 00:07:41,370 --> 00:07:47,400 also 1 over N factorial, sum over alpha. 87 00:07:47,400 --> 00:07:51,700 Epsilon alpha P alpha. 88 00:07:51,700 --> 00:07:55,380 Where epsilon alpha, again, is that sine factor 89 00:07:55,380 --> 00:07:57,040 for each permutation. 90 00:07:57,040 --> 00:08:01,020 So here are two operators, and we're 91 00:08:01,020 --> 00:08:05,010 going to try to prove that these operators are 92 00:08:05,010 --> 00:08:10,080 orthogonal projectors that take you 93 00:08:10,080 --> 00:08:14,460 to the subspace's symmetric states 94 00:08:14,460 --> 00:08:16,620 and anti-symmetric states. 95 00:08:16,620 --> 00:08:20,580 So that is our first goal. 96 00:08:20,580 --> 00:08:24,090 Understanding that these operators do the job. 97 00:08:24,090 --> 00:08:28,350 And then we'll see what we can do with them. 98 00:08:30,910 --> 00:08:32,640 So there are several things that we 99 00:08:32,640 --> 00:08:36,510 have to understand about these operators. 100 00:08:36,510 --> 00:08:41,850 If they are projectors, they should square to themselves. 101 00:08:41,850 --> 00:08:45,780 So S times S should be equal to S. Remember, that's 102 00:08:45,780 --> 00:08:48,030 the main equation of a projector. 103 00:08:48,030 --> 00:08:51,750 A projector is P squared equals P. 104 00:08:51,750 --> 00:08:56,500 And the second thing, they should be Hermitian. 105 00:08:56,500 --> 00:08:59,200 So let's try each of them. 106 00:08:59,200 --> 00:09:12,070 So the first claim is that S and A are Hermitian. 107 00:09:15,970 --> 00:09:28,170 In particular, S dagger equal S. And A dagger equal A. 108 00:09:28,170 --> 00:09:38,350 So if you want to prove something like that, 109 00:09:38,350 --> 00:09:42,450 it's not completely obvious at first sight 110 00:09:42,450 --> 00:09:45,120 that those statements are true. 111 00:09:45,120 --> 00:09:49,500 Because we saw, for example, that transpositions 112 00:09:49,500 --> 00:09:51,960 are Hermitian operators. 113 00:09:51,960 --> 00:09:57,980 But the general permutation operator is not Hermitian. 114 00:09:57,980 --> 00:09:59,070 It's unitary. 115 00:10:02,400 --> 00:10:03,690 So it's not so obvious. 116 00:10:03,690 --> 00:10:06,210 You cannot just say each operator is Hermitian, 117 00:10:06,210 --> 00:10:08,010 and it just works out. 118 00:10:08,010 --> 00:10:10,620 It's a little more complicated than that. 119 00:10:10,620 --> 00:10:17,230 But it's not extremely more complicated than that. 120 00:10:17,230 --> 00:10:21,180 So let's think of the following statement. 121 00:10:21,180 --> 00:10:25,410 I claim that if you have the list of all the permutation 122 00:10:25,410 --> 00:10:26,190 operators. 123 00:10:26,190 --> 00:10:27,790 Put the list in front of you. 124 00:10:27,790 --> 00:10:29,400 All of them. 125 00:10:29,400 --> 00:10:35,130 And you apply Hermitian conjugation to that whole list, 126 00:10:35,130 --> 00:10:38,950 you get another list of operators. 127 00:10:38,950 --> 00:10:42,960 And it will be just the same list scrambled. 128 00:10:42,960 --> 00:10:45,740 But you will get the same list. 129 00:10:45,740 --> 00:10:48,630 It will contain all of them. 130 00:10:48,630 --> 00:10:55,260 And that is kind of obvious if you think about it a little 131 00:10:55,260 --> 00:10:56,370 more. 132 00:10:56,370 --> 00:11:01,200 You have here, for example, the list of all the permutations. 133 00:11:01,200 --> 00:11:06,240 And here, you apply Hermitian conjugation, HC or dagger. 134 00:11:08,910 --> 00:11:10,340 And you get another list. 135 00:11:15,370 --> 00:11:20,830 And this list is the same as this one although reordered. 136 00:11:20,830 --> 00:11:24,640 In a sense, it permutes the permutation operators, 137 00:11:24,640 --> 00:11:27,250 if you wish. 138 00:11:27,250 --> 00:11:31,390 And the reason is clear. 139 00:11:31,390 --> 00:11:35,310 If you have two operators here, and you apply-- 140 00:11:35,310 --> 00:11:38,210 they are different-- and you apply Hermitian conjugation, 141 00:11:38,210 --> 00:11:41,350 it should give two different operators here. 142 00:11:41,350 --> 00:11:45,130 Because if they were equal, you could apply again 143 00:11:45,130 --> 00:11:47,410 Hermitian operator conjugation. 144 00:11:47,410 --> 00:11:49,060 And you would say, oh, they're equal. 145 00:11:49,060 --> 00:11:50,650 But you assume they were different. 146 00:11:50,650 --> 00:11:55,940 So two different operators go to two different operators here. 147 00:11:55,940 --> 00:11:58,840 Moreover, any operator here-- 148 00:11:58,840 --> 00:12:01,270 you can call it O-- 149 00:12:01,270 --> 00:12:05,950 is really equal to O dagger. 150 00:12:05,950 --> 00:12:06,700 O dagger. 151 00:12:06,700 --> 00:12:09,430 Hermitian conjugation twice gives you back. 152 00:12:09,430 --> 00:12:14,890 So any operator here is the Hermitian conjugation. 153 00:12:14,890 --> 00:12:18,500 Hermitian conjugate of some operator there. 154 00:12:18,500 --> 00:12:21,880 So every operator here will end up 155 00:12:21,880 --> 00:12:24,370 at the end of an arrow coming here. 156 00:12:24,370 --> 00:12:27,310 Maybe those arrows, we could put them 157 00:12:27,310 --> 00:12:31,040 like they go in all directions. 158 00:12:31,040 --> 00:12:33,370 But the fact is that when you take 159 00:12:33,370 --> 00:12:36,380 Hermitian conjugation of all the list, 160 00:12:36,380 --> 00:12:37,825 you get all the list back. 161 00:12:40,940 --> 00:12:44,690 Therefore, that takes care of the symmetrizer. 162 00:12:44,690 --> 00:12:48,410 When I take the Hermitian conjugation of P alpha, 163 00:12:48,410 --> 00:12:50,750 I might get another P beta. 164 00:12:50,750 --> 00:12:53,630 But at the end of the day, the whole sum 165 00:12:53,630 --> 00:12:56,360 will become again the whole sum. 166 00:12:56,360 --> 00:13:05,315 So that proves the fact that this map is 1 to 1 167 00:13:05,315 --> 00:13:12,060 The map Hermitian conjugation is 1 to 1. 168 00:13:12,060 --> 00:13:18,840 1 and surjective meaning that every element 169 00:13:18,840 --> 00:13:20,580 is reached as well. 170 00:13:20,580 --> 00:13:26,730 Means that S dagger is equal to S. 171 00:13:26,730 --> 00:13:30,750 Because the list doesn't change. 172 00:13:30,750 --> 00:13:37,200 Here is a slightly more subtle point. 173 00:13:37,200 --> 00:13:41,820 This time, the list is going to change. 174 00:13:41,820 --> 00:13:45,290 But then maybe when this changes to another operator, 175 00:13:45,290 --> 00:13:47,930 maybe the epsilon is not the right epsilon. 176 00:13:47,930 --> 00:13:51,630 You would have to worry about that. 177 00:13:51,630 --> 00:13:53,780 But that's no worry either. 178 00:13:53,780 --> 00:13:56,600 That also works out clearly. 179 00:13:56,600 --> 00:13:58,700 Because of the following thing. 180 00:13:58,700 --> 00:14:02,960 If a permutation P alpha is even, 181 00:14:02,960 --> 00:14:06,770 it's billed with an even number of transpositions. 182 00:14:06,770 --> 00:14:10,070 If you take the Hermitian conjugation 183 00:14:10,070 --> 00:14:12,320 of an even number of transpositions, 184 00:14:12,320 --> 00:14:14,720 you'll get an even number of transpositions 185 00:14:14,720 --> 00:14:16,680 in the other order. 186 00:14:16,680 --> 00:14:18,940 That's what Hermitian conjugation is. 187 00:14:18,940 --> 00:14:24,410 Therefore, Hermitian conjugation doesn't change the fact 188 00:14:24,410 --> 00:14:28,950 that the permutation is even or is odd. 189 00:14:28,950 --> 00:14:39,260 So if P alpha has a sine factor E of P alpha. 190 00:14:39,260 --> 00:14:41,870 Let's write it more explicitly. 191 00:14:41,870 --> 00:14:51,320 And P alpha dagger has as a sine factor E of P alpha dagger. 192 00:14:51,320 --> 00:14:52,960 These two are the same. 193 00:14:57,520 --> 00:15:00,330 If P alpha is even, P alpha dagger is even. 194 00:15:00,330 --> 00:15:03,690 If P alpha is odd, P alpha is odd. 195 00:15:03,690 --> 00:15:07,410 Therefore, when you take the Hermitian conjugate 196 00:15:07,410 --> 00:15:10,200 of this thing, you may map. 197 00:15:10,200 --> 00:15:13,380 You think the Hermitian conjugate of this sum. 198 00:15:13,380 --> 00:15:21,780 You have here a dagger would be 1 over N factorial, sum 199 00:15:21,780 --> 00:15:29,430 over alpha, epsilon alpha, P alpha dagger, which 200 00:15:29,430 --> 00:15:31,230 is another permutation. 201 00:15:31,230 --> 00:15:35,430 But this permutation. 202 00:15:35,430 --> 00:15:37,980 For this permutation, this sine factor 203 00:15:37,980 --> 00:15:39,660 is the correct sine factor. 204 00:15:39,660 --> 00:15:41,790 Because the sine factor for P alpha 205 00:15:41,790 --> 00:15:45,630 dagger is the same as the sine factor of E alpha. 206 00:15:45,630 --> 00:15:50,130 Since we get the whole sum by the statement before, 207 00:15:50,130 --> 00:15:54,050 A is also Hermitian. 208 00:15:54,050 --> 00:15:56,060 So OK. 209 00:15:56,060 --> 00:16:02,940 Not completely obvious, but that's a fact. 210 00:16:02,940 --> 00:16:06,740 If you think more abstractly, this 211 00:16:06,740 --> 00:16:09,500 is something you could know on a group. 212 00:16:09,500 --> 00:16:13,880 If you have a group that has all kinds of elements. 213 00:16:13,880 --> 00:16:19,430 Many of you have written papers in groups in your essays. 214 00:16:19,430 --> 00:16:22,160 Then if you take every group element, 215 00:16:22,160 --> 00:16:25,470 and you replace it by its inverse, 216 00:16:25,470 --> 00:16:27,950 you get the same list of group elements. 217 00:16:27,950 --> 00:16:29,750 It's just scrambled. 218 00:16:29,750 --> 00:16:35,450 And remember, since this permutations are unitary, 219 00:16:35,450 --> 00:16:38,590 taking Hermitian conjugation is the same thing 220 00:16:38,590 --> 00:16:40,710 as taking an inverse. 221 00:16:40,710 --> 00:16:42,860 So when you take a group, and you 222 00:16:42,860 --> 00:16:44,810 take the inverse of every element, 223 00:16:44,810 --> 00:16:47,860 you get back the list of the elements of the group. 224 00:16:47,860 --> 00:16:50,470 And that's what's happening here.