1
00:00:01,410 --> 00:00:06,920
PROFESSOR: So here comes the
point that this quite fabulous

2
00:00:06,920 --> 00:00:09,180
about Hermitian operators.

3
00:00:09,180 --> 00:00:15,700
Here is the thing that it
really should impress you.

4
00:00:15,700 --> 00:00:22,880
It's the fact that any,
all Hermitian operators

5
00:00:22,880 --> 00:00:29,335
have as many eigenfunctions and
eigenvalues as you can possibly

6
00:00:29,335 --> 00:00:33,470
need, whatever that means.

7
00:00:33,470 --> 00:00:35,720
But they're rich.

8
00:00:35,720 --> 00:00:41,020
It's a lot of those states.

9
00:00:41,020 --> 00:00:49,300
What it really means is that the
set of eigenfunctions for any

10
00:00:49,300 --> 00:00:52,270
Hermitian operator--
whatever Hermitian operator,

11
00:00:52,270 --> 00:00:55,300
it's not just for some
especially nice ones--

12
00:00:55,300 --> 00:01:01,130
for all of them you
get eigenfunctions.

13
00:01:01,130 --> 00:01:06,490
And these eigenfunctions,
because it has vectors,

14
00:01:06,490 --> 00:01:11,230
they are enough to span
the space of states.

15
00:01:11,230 --> 00:01:15,130
That is any state can be
written as a superposition

16
00:01:15,130 --> 00:01:16,940
of those eigenvectors.

17
00:01:16,940 --> 00:01:18,120
There's enough.

18
00:01:18,120 --> 00:01:24,270
If you're thinking finite
dimensional vector spaces,

19
00:01:24,270 --> 00:01:28,060
if you're looking at
the Hermitian matrix,

20
00:01:28,060 --> 00:01:31,120
the eigenvectors
will provide you

21
00:01:31,120 --> 00:01:34,120
a basis for the vector space.

22
00:01:34,120 --> 00:01:38,340
You can understand anything
in terms of eigenvectors.

23
00:01:38,340 --> 00:01:40,870
It is such an important theorem.

24
00:01:40,870 --> 00:01:45,330
It's called the spectral
theorem in mathematics.

25
00:01:45,330 --> 00:01:49,870
And it's discussed in
lots of detail in 805.

26
00:01:49,870 --> 00:01:52,480
Because there's
a minor subtlety.

27
00:01:52,480 --> 00:01:56,190
We can get the whole
idea about it here.

28
00:01:56,190 --> 00:01:58,750
But there are a couple
of complications

29
00:01:58,750 --> 00:02:00,940
that mathematicians
have to iron out.

30
00:02:00,940 --> 00:02:06,100
So basically let's
state we really

31
00:02:06,100 --> 00:02:08,370
need, which is the following.

32
00:02:08,370 --> 00:02:26,581
Consider the collection of
eigenfunctions and eigenvalues

33
00:02:26,581 --> 00:02:34,130
of the Hermitian operator q.

34
00:02:34,130 --> 00:02:43,380
And then I go and say, well,
q psi 1 equal q 1 psi 1 q psi

35
00:02:43,380 --> 00:02:45,620
2 equal q2 psi 2.

36
00:02:50,458 --> 00:02:54,620
And I actually don't
specify if it's

37
00:02:54,620 --> 00:02:58,500
a finite set or an infinite set.

38
00:02:58,500 --> 00:03:02,090
The infinite set, of course,
is a tiny bit more complicated.

39
00:03:02,090 --> 00:03:05,282
But the result is true as well.

40
00:03:05,282 --> 00:03:07,350
And we can work with it.

41
00:03:07,350 --> 00:03:12,590
So that is the set up.

42
00:03:12,590 --> 00:03:14,005
And here comes the claim.

43
00:03:18,050 --> 00:03:32,470
Claim 3, the
eigenfunctions can be

44
00:03:32,470 --> 00:03:46,290
organized to satisfy the
following relation, integral dx

45
00:03:46,290 --> 00:03:55,870
psi i of x psi j of x
is equal to delta ij.

46
00:03:58,510 --> 00:04:00,830
And this is called
orthonormality.

47
00:04:06,620 --> 00:04:08,825
Let's see what this all means.

48
00:04:11,450 --> 00:04:13,890
We have a collection
of eigenfunctions.

49
00:04:13,890 --> 00:04:18,110
And here it says
something quite nice.

50
00:04:29,090 --> 00:04:35,610
These functions are like
orthonormal functions,

51
00:04:35,610 --> 00:04:40,900
which is to say each
function has unit norm.

52
00:04:40,900 --> 00:04:45,430
You see, if you
take i equal to j,

53
00:04:45,430 --> 00:04:51,470
suppose you take psi 1 psi 1,
you get delta 1 1, which is 1.

54
00:04:51,470 --> 00:04:55,140
Remember the [INAUDIBLE] for
delta is 1 from the [INAUDIBLE]

55
00:04:55,140 --> 00:04:56,020
are the same.

56
00:04:56,020 --> 00:05:01,600
And it's 0 otherwise.
psi 1 the norm of psi 1

57
00:05:01,600 --> 00:05:04,660
is 1 and [INAUDIBLE]
squared [INAUDIBLE]

58
00:05:04,660 --> 00:05:09,090
psi 1, psi 2, psi3, all of
them are well normalized.

59
00:05:09,090 --> 00:05:14,220
So they satisfied this thing
we wanted them to satisfy.

60
00:05:14,220 --> 00:05:17,500
Those are good states.

61
00:05:17,500 --> 00:05:20,020
psi 1, psi 2, psi 3,
those are good states.

62
00:05:20,020 --> 00:05:21,840
They are all normalized.

63
00:05:21,840 --> 00:05:29,410
But even more, any two
different ones are orthonormal.

64
00:05:29,410 --> 00:05:34,470
This is like the 3
basis vectors of r3.

65
00:05:34,470 --> 00:05:39,880
The x basic unit vector, the y
unit vector, the z unit vector,

66
00:05:39,880 --> 00:05:43,880
each one has length 1, and
they're all orthonormal.

67
00:05:43,880 --> 00:05:46,840
And when are two
functions orthonormal?

68
00:05:46,840 --> 00:05:50,900
You say, well, when vectors are
orthonormal I know what I mean.

69
00:05:50,900 --> 00:05:55,870
But orthonormality for functions
means doing this integral.

70
00:05:55,870 --> 00:06:00,840
This measures how different one
function is from another one.

71
00:06:00,840 --> 00:06:02,890
Because if you have
the same function,

72
00:06:02,890 --> 00:06:06,935
this integral and this
positive, and this all adds up.

73
00:06:06,935 --> 00:06:10,030
But for different
functions, this

74
00:06:10,030 --> 00:06:15,290
is a measure of the inner
product between two functions.

75
00:06:15,290 --> 00:06:18,300
You see, you have the dot
product between two vectors.

76
00:06:18,300 --> 00:06:22,675
The dot product of two functions
is an integral like that.

77
00:06:22,675 --> 00:06:29,130
It's the only thing
that makes sense

78
00:06:29,130 --> 00:06:33,130
So I want to prove
one part of this,

79
00:06:33,130 --> 00:06:38,030
which is a part that is doable
with elementary methods.

80
00:06:38,030 --> 00:06:44,450
And the other part is a
little more complicated.

81
00:06:44,450 --> 00:06:48,620
So let's do this.

82
00:06:48,620 --> 00:06:59,400
And consider the case if
qi is different from qj,

83
00:06:59,400 --> 00:07:03,460
I claim i can prove
this property.

84
00:07:03,460 --> 00:07:06,250
We can prove this
orthonormality.

85
00:07:06,250 --> 00:07:17,320
So start with the integral
dx of psi i star q psi j.

86
00:07:22,070 --> 00:07:27,460
Well, q out here at psi j is qj.

87
00:07:27,460 --> 00:07:38,230
So this is integral dx
psi i star qj psi j.

88
00:07:38,230 --> 00:07:44,772
And therefore, it's equal to qj
times integral psi i star psi

89
00:07:44,772 --> 00:07:46,100
j.

90
00:07:46,100 --> 00:07:53,880
I simplified this by
just enervating it.

91
00:07:53,880 --> 00:07:58,310
Because psi i and psi
j are eigenstates of q.

92
00:07:58,310 --> 00:08:01,760
Now, the other thing I
can do is use the property

93
00:08:01,760 --> 00:08:06,890
that q is Hermitant and move
the q to act on this function.

94
00:08:06,890 --> 00:08:18,270
So this is equal to integral
dx q i psi i star psi j.

95
00:08:18,270 --> 00:08:22,390
And now I can keep
simplifying as well.

96
00:08:22,390 --> 00:08:23,420
And I have dx.

97
00:08:23,420 --> 00:08:33,081
And then I have the complex
conjugate of qi psi i psi i,

98
00:08:33,081 --> 00:08:37,970
like this, psi j.

99
00:08:37,970 --> 00:08:42,490
And now, remember q is an
eigenvalue for Hermitian

100
00:08:42,490 --> 00:08:43,570
operator.

101
00:08:43,570 --> 00:08:45,320
We already know it's real.

102
00:08:45,320 --> 00:08:49,890
So q goes out of the
integral as a number.

103
00:08:49,890 --> 00:08:53,080
Because it's real,
and it's not changed.

104
00:08:53,080 --> 00:08:59,100
Integral dx psi i star psi j.

105
00:08:59,100 --> 00:09:03,600
The end result is that we've
shown that this quantity is

106
00:09:03,600 --> 00:09:07,251
equal to this second quantity.

107
00:09:07,251 --> 00:09:10,030
And therefore moving this--

108
00:09:10,030 --> 00:09:13,850
since the integral is the same
in both quantities, this shows

109
00:09:13,850 --> 00:09:21,900
that q i minus qj, subtracting
these two equations,

110
00:09:21,900 --> 00:09:31,530
or just moving one to one side,
integral psi i star psi j dx

111
00:09:31,530 --> 00:09:34,550
is equal to 0.

112
00:09:34,550 --> 00:09:37,350
So look what you've proven
by using Hermiticity,

113
00:09:37,350 --> 00:09:41,410
that the difference
between the eigenvalues

114
00:09:41,410 --> 00:09:47,520
times the overlap between
psi i and psi j must be 0.

115
00:09:47,520 --> 00:09:49,800
But we started
with the assumption

116
00:09:49,800 --> 00:09:52,970
that the eigenvalues
are different.

117
00:09:52,970 --> 00:09:55,780
And if the eigenvalues are
different, this is non-zero.

118
00:09:55,780 --> 00:09:59,810
And the only possibility
is that this integral is 0.

119
00:09:59,810 --> 00:10:03,190
So this implies
since we've assumed

120
00:10:03,190 --> 00:10:08,310
that qi is different than qj.

121
00:10:08,310 --> 00:10:17,740
We've proven that psi i
star psi j dx is equal to 0.

122
00:10:17,740 --> 00:10:20,960
And that's part of
this little theorem.

123
00:10:20,960 --> 00:10:24,740
That the eigenfunctions
can be organized

124
00:10:24,740 --> 00:10:28,580
to have orthonormality
and orthonormality

125
00:10:28,580 --> 00:10:31,510
between the different points.

126
00:10:31,510 --> 00:10:32,980
My proof is good.

127
00:10:32,980 --> 00:10:36,470
But it's not perfect.

128
00:10:36,470 --> 00:10:40,750
Because it ignores one
possible complication,

129
00:10:40,750 --> 00:10:47,580
which is that here we wrote the
list of all the eigenfunctions.

130
00:10:47,580 --> 00:10:50,550
But sometimes something
very interesting

131
00:10:50,550 --> 00:10:52,770
happens in quantum mechanics.

132
00:10:52,770 --> 00:10:55,490
It's called degeneracy.

133
00:10:55,490 --> 00:10:58,710
And degeneracy
means that there may

134
00:10:58,710 --> 00:11:05,220
be several eigenfunctions
that are different but have

135
00:11:05,220 --> 00:11:06,766
the same eigenvalue.

136
00:11:09,940 --> 00:11:12,210
We're going to find that soon--

137
00:11:12,210 --> 00:11:14,220
we're going to find,
for example, states

138
00:11:14,220 --> 00:11:18,880
of a particle that move in
a circle that are different

139
00:11:18,880 --> 00:11:20,260
and have the same energy.

140
00:11:20,260 --> 00:11:23,560
For example, a particle moving
in a circle with this velocity

141
00:11:23,560 --> 00:11:27,530
and a particle moving in a
circle with the same magnitude

142
00:11:27,530 --> 00:11:30,440
of the velocity in
the other direction

143
00:11:30,440 --> 00:11:35,250
are two states that
are different but have

144
00:11:35,250 --> 00:11:39,040
the same energy eigenvalue.

145
00:11:39,040 --> 00:11:42,440
So it's possible that this
list not all are different.

146
00:11:42,440 --> 00:11:47,050
So suppose you have like three
or four degenerate states,

147
00:11:47,050 --> 00:11:49,420
say three degenerate states.

148
00:11:49,420 --> 00:11:51,930
They all have the
same eigenvalue.

149
00:11:51,930 --> 00:11:53,330
But they are different.

150
00:11:53,330 --> 00:11:57,070
Are they orthonormal or not?

151
00:11:57,070 --> 00:12:06,046
The answer is-- actually
the clue is there.

152
00:12:06,046 --> 00:12:11,940
The eigenfunctions can
be organized to satisfy.

153
00:12:11,940 --> 00:12:16,340
It would be wrong if you say
the eigenfunctions satisfy.

154
00:12:16,340 --> 00:12:19,260
They can be
organized to satisfy.

155
00:12:19,260 --> 00:12:25,620
It means that, yes, those ones
that have different eigenvalues

156
00:12:25,620 --> 00:12:28,063
are automatically orthonormal.

157
00:12:28,063 --> 00:12:32,130
But those that have
the same eigenvalues,

158
00:12:32,130 --> 00:12:35,440
you may have three
of them maybe,

159
00:12:35,440 --> 00:12:37,990
they may not necessarily
be orthonormal.

160
00:12:37,990 --> 00:12:43,490
But you can do linear
transformations of them

161
00:12:43,490 --> 00:12:48,010
and form linear combinations
such that they are orthonormal.

162
00:12:48,010 --> 00:12:51,070
So the interesting
part of this theorem,

163
00:12:51,070 --> 00:12:54,070
which is the more difficult
part mathematically,

164
00:12:54,070 --> 00:12:57,370
is to show that when
you have degeneracies

165
00:12:57,370 --> 00:12:59,750
this still can be done.

166
00:12:59,750 --> 00:13:06,100
And there's still enough
eigenvectors to span the space.