1
00:00:06 --> 00:00:08
OK, this is lecture twenty.

2
00:00:08 --> 00:00:13
And this is the final lecture
on determinants.

3
00:00:13 --> 00:00:16
And it's about the
applications.

4
00:00:16 --> 00:00:21
So we worked hard in the last
two lectures to get a formula

5
00:00:21 --> 00:00:26
for the determinant and the
properties of the determinant.

6
00:00:26 --> 00:00:32
Now to use the determinant and,
and always this determinant

7
00:00:32 --> 00:00:36
packs
all this information into a

8
00:00:36 --> 00:00:38
single number.

9
00:00:38 --> 00:00:44
And that number can give us
formulas for all sorts of,

10
00:00:44 --> 00:00:49
things that we've been
calculating already without

11
00:00:49 --> 00:00:50
formulas.

12
00:00:50 --> 00:00:52
Now what was A inverse?

13
00:00:52 --> 00:00:58.42
So, so I'm beginning with the
formula for A inverse.

14
00:00:58.42 --> 00:01:03
Two, two by two formula we
know,

15
00:01:03 --> 00:01:03
right?

16
00:01:03 --> 00:01:10
The two by two formula for A
inverse, the inverse of a b c d

17
00:01:10 --> 00:01:17
inverse is one over the
determinant times d a -b -c.

18
00:01:17 --> 00:01:24
Somehow I want to see what's
going on for three by three and

19
00:01:24 --> 00:01:25
n by n.

20
00:01:25 --> 00:01:32
And actually maybe you can see
what's going on from this two by

21
00:01:32 --> 00:01:33
two case.

22
00:01:33 --> 00:01:39
So there's a formula
for the inverse,

23
00:01:39 --> 00:01:41
and what did I divide by?

24
00:01:41 --> 00:01:43
The determinant.

25
00:01:43 --> 00:01:48
So what I'm looking for is a
formula where it has one over

26
00:01:48 --> 00:01:53
the determinant and,
and you remember why that makes

27
00:01:53 --> 00:01:58
good sense, because then that's
perfect as long as the

28
00:01:58 --> 00:02:03
determinant isn't zero,
and that's exactly when there

29
00:02:03 --> 00:02:05
is an
inverse.

30
00:02:05 --> 00:02:10
But now I have to ask can we
recognize any of this stuff?

31
00:02:10 --> 00:02:14.9
Do you recognize what that
number d is from the past?

32
00:02:14.9 --> 00:02:17
From last, from the last
lecture?

33
00:02:17 --> 00:02:19
My hint is think cofactors.

34
00:02:19 --> 00:02:24
Because my formula is going to
be, my formula for the inverse

35
00:02:24 --> 00:02:29
is going to be one over the
determinant times

36
00:02:29 --> 00:02:31
a matrix of cofactors.

37
00:02:31 --> 00:02:33
So you remember that D?

38
00:02:33 --> 00:02:36
What's that the cofactor of?

39
00:02:36 --> 00:02:38
Remember cofactors?

40
00:02:38 --> 00:02:43
If -- that's the one one
cofactor, because if I strike

41
00:02:43 --> 00:02:47
out row and column one,
I'm left with d.

42
00:02:47 --> 00:02:49
And what's minus b?

43
00:02:49 --> 00:02:49
OK.

44
00:02:49 --> 00:02:52
Which cofactor is that one?

45
00:02:52 --> 00:02:56.65
Oh, minus b is the cofactor of
c,

46
00:02:56.65 --> 00:02:57
right?

47
00:02:57 --> 00:03:02
If I strike out the c,
I'm left with a b there.

48
00:03:02 --> 00:03:04
And why the minus sign?

49
00:03:04 --> 00:03:11
Because this c was in a two one
position, and two plus one is

50
00:03:11 --> 00:03:11
odd.

51
00:03:11 --> 00:03:16
So there was a minus went into
the cofactor,

52
00:03:16 --> 00:03:17
and that's it.

53
00:03:17 --> 00:03:17
OK.

54
00:03:17 --> 00:03:23
I'll write down next what my
formula is.

55
00:03:23 --> 00:03:28
Here's the big formula for the
A -- for A inverse.

56
00:03:28 --> 00:03:34
It's one over the determinant
of A and then some matrix.

57
00:03:34 --> 00:03:39
And that matrix is the matrix
of cofactors,

58
00:03:39 --> 00:03:39
c.

59
00:03:39 --> 00:03:43
Only one thing,
it turns -- you'll see,

60
00:03:43 --> 00:03:45
I have to, I transpose.

61
00:03:45 --> 00:03:50
So this is the matrix of
cofactors,

62
00:03:50 --> 00:03:55
the -- what I'll just -- but
why don't we just call it the

63
00:03:55 --> 00:03:57
cofactor matrix.

64
00:03:57 --> 00:04:01
So the one one entry of,
of c is the cof- is the one one

65
00:04:01 --> 00:04:07
cofactor, the thing that we get
by throwing away row and column

66
00:04:07 --> 00:04:07
one.

67
00:04:07 --> 00:04:08
It's the d.

68
00:04:08 --> 00:04:14
And, because of the transpose,
what I see up here is the

69
00:04:14 --> 00:04:17
cofactor of this guy down here,
right?

70
00:04:17 --> 00:04:20
That's where the transpose came
in.

71
00:04:20 --> 00:04:23
What I see here,
this is the cofactor not of

72
00:04:23 --> 00:04:26
this one, because I've
transposed.

73
00:04:26 --> 00:04:28
This is the cofactor of the b.

74
00:04:28 --> 00:04:34
When I throw away the b,
the b row and the b column,

75
00:04:34 --> 00:04:38
I'm left with c,
and then I have that minus sign

76
00:04:38 --> 00:04:38
again.

77
00:04:38 --> 00:04:43
And of course the two two entry
is the cofactor of d,

78
00:04:43 --> 00:04:44.59
and that's this a.

79
00:04:44.59 --> 00:04:44
OK.

80
00:04:44 --> 00:04:46
So there's the formula.

81
00:04:46 --> 00:04:48
But we got to think why.

82
00:04:48 --> 00:04:53
I mean, it worked in this two
by two case, but a lot of other

83
00:04:53 --> 00:04:58
formulas would
have worked just as well.

84
00:04:58 --> 00:05:01
We, we have to see why that's
true.

85
00:05:01 --> 00:05:05
In other words,
why is it -- so this is what I

86
00:05:05 --> 00:05:06
aim to find.

87
00:05:06 --> 00:05:12
And, and let's just sort of
look to see what is that telling

88
00:05:12 --> 00:05:12
us.

89
00:05:12 --> 00:05:18
That tells us that the --
the expression for A inverse --

90
00:05:18 --> 00:05:20.73
let's look at a three by three.

91
00:05:20.73 --> 00:05:24
Can I just write down a a b c d
e f g h i?

92
00:05:24 --> 00:05:26
And I'm looking for its
inverse.

93
00:05:26 --> 00:05:30
And what kind of a formula --
do I see there?

94
00:05:30 --> 00:05:35
I mean, what -- the determinant
is a bunch of products of three

95
00:05:35 --> 00:05:37.01
factors, right?

96
00:05:37.01 --> 00:05:40
The determinant of this
matrix'll

97
00:05:40 --> 00:05:43
involve a e i,
and b f times g,

98
00:05:43 --> 00:05:47.86
and c times d times h,
and minus c e g,

99
00:05:47.86 --> 00:05:48
and so on.

100
00:05:48 --> 00:05:52
So things with three factors go
in here.

101
00:05:52 --> 00:05:59
Things with how many factors do
things in the cofactor matrix

102
00:05:59 --> 00:05:59
have?

103
00:05:59 --> 00:06:02
What's a typical cofactor?

104
00:06:02 --> 00:06:05
What's
the cofactor of a?

105
00:06:05 --> 00:06:10
The cofactor of a,
the one one entry up here in

106
00:06:10 --> 00:06:11
the inverse is?

107
00:06:11 --> 00:06:17
I throw away the row and column
containing a and I take the

108
00:06:17 --> 00:06:22.12
determinant of what's left,
that's the cofactor.

109
00:06:22.12 --> 00:06:24
And that's e i minus f h.

110
00:06:24 --> 00:06:26
Products of two things.

111
00:06:26 --> 00:06:31
Now,
I'm just making the observation

112
00:06:31 --> 00:06:38
that the determinant of A
involves products of n entries.

113
00:06:38 --> 00:06:44
And the cofactor matrix
involves products of n minus 1

114
00:06:44 --> 00:06:45
entries.

115
00:06:45 --> 00:06:52
And, like, we never noticed any
of this stuff when we were

116
00:06:52 --> 00:07:00
computing the inverse by the
Gauss-Jordan method or whatever.

117
00:07:00 --> 00:07:02.58
You remember how we did it?

118
00:07:02.58 --> 00:07:06
We took the matrix A,
we tucked the identity next to

119
00:07:06 --> 00:07:09
it, we did elimination till A
became the identity.

120
00:07:09 --> 00:07:12
And then that,
the identity suddenly was A

121
00:07:12 --> 00:07:13
inverse.

122
00:07:13 --> 00:07:15
Well, that was great
numerically.

123
00:07:15 --> 00:07:19
But we never knew what was
going on, basically.

124
00:07:19 --> 00:07:22
And now we see what the formula
is

125
00:07:22 --> 00:07:27
in terms of letters,
what's the algebra instead of

126
00:07:27 --> 00:07:28
the algorithm.

127
00:07:28 --> 00:07:28
OK.

128
00:07:28 --> 00:07:32
But I have to say why this is
right, right?

129
00:07:32 --> 00:07:36
I still -- that's a pretty
magic formula.

130
00:07:36 --> 00:07:38
Where does it come from?

131
00:07:38 --> 00:07:40
Well, I'll just check it.

132
00:07:40 --> 00:07:44
Having,
having got it up there,

133
00:07:44 --> 00:07:47
let me -- I'll say,
how can we check -- what do I

134
00:07:47 --> 00:07:48
want to check?

135
00:07:48 --> 00:07:53
I want to check that A times
its inverse gives the identity.

136
00:07:53 --> 00:07:57
So I want, I want to check that
A times this thing,

137
00:07:57 --> 00:08:01
A times this -- now I'm going
to write in the inverse -- gives

138
00:08:01 --> 00:08:03
the identity.

139
00:08:03 --> 00:08:08
So I check that A times C
transpose -- let me bring the

140
00:08:08 --> 00:08:10
determinant up here.

141
00:08:10 --> 00:08:14
Determinant of A times the
identity.

142
00:08:14 --> 00:08:15
That's my job.

143
00:08:15 --> 00:08:20.02
That's it, that if this is
true, and it is,

144
00:08:20.02 --> 00:08:25
then, then I've correctly
identified A inverse as C

145
00:08:25 --> 00:08:29
transpose divided by the
determinant.

146
00:08:29 --> 00:08:30
OK.

147
00:08:30 --> 00:08:32
But why is this true?

148
00:08:32 --> 00:08:33
Why is that true?

149
00:08:33 --> 00:08:37
Let me, let me put down what
I'm doing here.

150
00:08:37 --> 00:08:42
I have A -- here,
here's A, here's a11 -- I'm

151
00:08:42 --> 00:08:45
doing this multiplication --
along to a1n.

152
00:08:45 --> 00:08:51
And then down in this last row
will be an an1 along to ann.

153
00:08:51 --> 00:08:57
And I'm multiplying by the
cofactor matrix transposed.

154
00:08:57 --> 00:09:02
So when I transpose,
it'll be c11 c12 down to c1n.

155
00:09:02 --> 00:09:08
Notice usually that one coming
first would mean I'm in row one,

156
00:09:08 --> 00:09:13
but I've transposed,
so that's, those are the

157
00:09:13 --> 00:09:14
cofactors.

158
00:09:14 --> 00:09:18
This first column are the
cofactors from row one.

159
00:09:18 --> 00:09:25
And then the last column would
be the cofactors from row n.

160
00:09:25 --> 00:09:30
And why should that come out to
be anything good?

161
00:09:30 --> 00:09:35
In fact, why should it come out
to be as good as this?

162
00:09:35 --> 00:09:41
Well, you can tell me what the
one one entry in the product is.

163
00:09:41 --> 00:09:47
This is like you're seeing the
main point if you just tell me

164
00:09:47 --> 00:09:48
one entry.

165
00:09:48 --> 00:09:52
What do I get up there in the
one

166
00:09:52 --> 00:09:58
one entry when I do this row of
this row from A times this

167
00:09:58 --> 00:10:00
column of cofactors?

168
00:10:00 --> 00:10:03
What, what will I get there?

169
00:10:03 --> 00:10:06
Because we have seen this.

170
00:10:06 --> 00:10:10
I mean, we're,
right, building exactly on what

171
00:10:10 --> 00:10:13
the last lecture reached.

172
00:10:13 --> 00:10:18
So this is a11 times c11,
a12 times c12,

173
00:10:18 --> 00:10:19.76
a1n times c1n.

174
00:10:19.76 --> 00:10:24
What does that what does that
sum up to?

175
00:10:24 --> 00:10:29.34
That's the cofactor formula for
the determinant.

176
00:10:29.34 --> 00:10:34
That's the, this cofactor
formula, which I wrote,

177
00:10:34 --> 00:10:37
which we got last time.

178
00:10:37 --> 00:10:41
The determinant of A is,
if I use row one,

179
00:10:41 --> 00:10:45
let,
let I equal one,

180
00:10:45 --> 00:10:49
then I have a11 times its
cofactor, a12 times its

181
00:10:49 --> 00:10:51
cofactor, and so on.

182
00:10:51 --> 00:10:55
And that gives me the
determinant.

183
00:10:55 --> 00:10:58
And it worked in this,
in this case.

184
00:10:58 --> 00:11:02
This row times this thing is,
sure enough,

185
00:11:02 --> 00:11:03
ad minus bc.

186
00:11:03 --> 00:11:08
But this formula says it always
works.

187
00:11:08 --> 00:11:11
So up here in this,
in this position,

188
00:11:11 --> 00:11:14
I'm getting determinant of A.

189
00:11:14 --> 00:11:17
What about in the two two
position?

190
00:11:17 --> 00:11:21
Row two times column two there,
what, what is that?

191
00:11:21 --> 00:11:25
That's just the cofactors,
that's just row two times its

192
00:11:25 --> 00:11:26
cofactors.

193
00:11:26 --> 00:11:30
So of course I get the
determinant

194
00:11:30 --> 00:11:31.18
again.

195
00:11:31.18 --> 00:11:35.9
And in the last here,
this is the last row times its

196
00:11:35.9 --> 00:11:36
cofactors.

197
00:11:36 --> 00:11:41
It's exactly -- you see,
we're realizing that the

198
00:11:41 --> 00:11:45
cofactor formula is just this
sum of products,

199
00:11:45 --> 00:11:49
so of course we think,
hey, we've got a matrix

200
00:11:49 --> 00:11:52
multiplication there.

201
00:11:52 --> 00:11:55.06
And we get determinant of A.

202
00:11:55.06 --> 00:11:55
Great.

203
00:11:55 --> 00:11:58
But there's one more idea here,
right?

204
00:11:58 --> 00:12:04
What else, what have I not --
so I haven't got that formula

205
00:12:04 --> 00:12:08
completely proved yet,
because I've still got to do

206
00:12:08 --> 00:12:13
all the off-diagonal stuff,
which I want to be zero,

207
00:12:13 --> 00:12:13
right?

208
00:12:13 --> 00:12:19
I just want this to be
determinant of A times the

209
00:12:19 --> 00:12:22
identity, and then I'm,
I'm a happy person.

210
00:12:22 --> 00:12:24
So why should that be?

211
00:12:24 --> 00:12:29
Why should it be that one row
times the cofactors from a

212
00:12:29 --> 00:12:33
different row,
which become a column because I

213
00:12:33 --> 00:12:35
transpose, give zero?

214
00:12:35 --> 00:12:39
In other words,
the cofactor formula gives the

215
00:12:39 --> 00:12:43
determinant if the
row and the,

216
00:12:43 --> 00:12:48
and the cofactors -- you know,
if the entries of A and the

217
00:12:48 --> 00:12:51
cofactors are for the same row.

218
00:12:51 --> 00:12:55
But for some reason,
if I take the cofactors from

219
00:12:55 --> 00:13:00
the -- entries from the first
row and the cofactors from the

220
00:13:00 --> 00:13:05
second row,
for some reason I automatically

221
00:13:05 --> 00:13:05
get zero.

222
00:13:05 --> 00:13:09
And it's sort of like
interesting to say,

223
00:13:09 --> 00:13:10
why does that happen?

224
00:13:10 --> 00:13:15
And can I just check that -- of
course, we know it happens,

225
00:13:15 --> 00:13:16
in this case.

226
00:13:16 --> 00:13:21
Here are the numbers from row
one and here are the cofactors

227
00:13:21 --> 00:13:23
from row two,
right?

228
00:13:23 --> 00:13:26
Those are the numbers in row
one.

229
00:13:26 --> 00:13:29
And th- these are the cofactors
from row two,

230
00:13:29 --> 00:13:34
because the cofactor of c is
minus b and the cofactor of d is

231
00:13:34 --> 00:13:34
a.

232
00:13:34 --> 00:13:38
And sure enough,
that row times this column

233
00:13:38 --> 00:13:39
gives -- please say it.

234
00:13:39 --> 00:13:40
Zero, right.

235
00:13:40 --> 00:13:41
OK.

236
00:13:41 --> 00:13:42
So now how come?

237
00:13:42 --> 00:13:43
How come?

238
00:13:43 --> 00:13:47
Can we even see it in
this two by two case?

239
00:13:47 --> 00:13:50
Why did -- well,
I mean, I guess we,

240
00:13:50 --> 00:13:53
you know, in one way we
certainly do see it,

241
00:13:53 --> 00:13:55
because it's right here.

242
00:13:55 --> 00:13:58
I mean, do we just do it,
and then we get zero.

243
00:13:58 --> 00:14:03
But we want to think of some
reason why the answer's zero,

244
00:14:03 --> 00:14:07
some reason that we can use in
the n by n case.

245
00:14:07 --> 00:14:10
So let -- here,
here is my thinking.

246
00:14:10 --> 00:14:14
We must be, if we're getting
the answer's zero,

247
00:14:14 --> 00:14:18
we suspect that what we're
doing somehow,

248
00:14:18 --> 00:14:23
we're taking the determinant of
some matrix that has two equal

249
00:14:23 --> 00:14:24
rows.

250
00:14:24 --> 00:14:29
So I believe that if we
multiply these by the cofactors

251
00:14:29 --> 00:14:34
from some other row,
we're taking the determinant --

252
00:14:34 --> 00:14:38
ye, what matrix are we taking
the determinant of?

253
00:14:38 --> 00:14:40
Here it's, this is it.

254
00:14:40 --> 00:14:45
We're, when we do that times
this, we're really taking -- can

255
00:14:45 --> 00:14:49
I put this in little letters
down here?

256
00:14:49 --> 00:14:54.47
I'm taking -- let me look at
the matrix a b a b.

257
00:14:54.47 --> 00:14:58
Let me call that matrix AS,
A screwed up.

258
00:14:58 --> 00:14:58
OK.

259
00:14:58 --> 00:14:59
All right.

260
00:14:59 --> 00:15:03
So now that matrix is certainly
singular.

261
00:15:03 --> 00:15:08
So if we find its determinant,
we're going to get zero.

262
00:15:08 --> 00:15:14
But I claim that if we find its
determinant by the cofactor

263
00:15:14 --> 00:15:18
rule,
go along the first row,

264
00:15:18 --> 00:15:22
we would take a times the
cofactor of a.

265
00:15:22 --> 00:15:28
And what is the -- see,
how -- oh no -- let me go along

266
00:15:28 --> 00:15:29
the second row.

267
00:15:29 --> 00:15:30.05
OK.

268
00:15:30.05 --> 00:15:34
So let's see,
which -- if I take -- I know

269
00:15:34 --> 00:15:37
I've got a singular matrix here.

270
00:15:37 --> 00:15:43
And I believe that when I do
this multiplication,

271
00:15:43 --> 00:15:48
what I'm doing is using the
cofactor formula for the

272
00:15:48 --> 00:15:49
determinant.

273
00:15:49 --> 00:15:52
And I know I'm going to get
zero.

274
00:15:52 --> 00:15:55
Let me try this again.

275
00:15:55 --> 00:16:01.07
So the cofactor formula for the
determinant says I should take a

276
00:16:01.07 --> 00:16:04.4
times its cofactor,
which is this b,

277
00:16:04.4 --> 00:16:10
plus b times its cofactor,
which is this minus a.

278
00:16:10 --> 00:16:10
OK.

279
00:16:10 --> 00:16:14
That's what we're doing,
apart from a sign here.

280
00:16:14 --> 00:16:18
Oh yeah, so you know,
there might be a minus

281
00:16:18 --> 00:16:20
multiplying everything.

282
00:16:20 --> 00:16:25
So if I take this determinant,
it's A -- the determinant of

283
00:16:25 --> 00:16:29
this, the determinant of A
screwed up is a times its

284
00:16:29 --> 00:16:33
cofactor, which is b,
plus the second guy

285
00:16:33 --> 00:16:37
times its cofactor,
which is minus a.

286
00:16:37 --> 00:16:43
And of course I get the answer
zero, and this is exactly what's

287
00:16:43 --> 00:16:47.93
happening in that,
in that, row times this wrong

288
00:16:47.93 --> 00:16:48
column.

289
00:16:48 --> 00:16:48
OK.

290
00:16:48 --> 00:16:54.47
That's the two by two picture,
and it's just the same here.

291
00:16:54.47 --> 00:16:59
That the reason I get a zero up
in there

292
00:16:59 --> 00:17:04
is, the reason I get a zero is
that when I multiply the first

293
00:17:04 --> 00:17:09
row of A and the last row of the
cofactor matrix,

294
00:17:09 --> 00:17:14
it's as if I'm taking the
determinant of this screwed up

295
00:17:14 --> 00:17:18
matrix that has first and last
rows identical.

296
00:17:18 --> 00:17:24
The book pins this down more
specific -- and more carefully

297
00:17:24 --> 00:17:27
than I can do in the lecture.

298
00:17:27 --> 00:17:30
I hope you're seeing the point.

299
00:17:30 --> 00:17:32
That this is an identity.

300
00:17:32 --> 00:17:36
That it's a beautiful identity
and it tells us what the inverse

301
00:17:36 --> 00:17:38
of the matrix is.

302
00:17:38 --> 00:17:42
So it gives us the inverse,
the formula for the inverse.

303
00:17:42 --> 00:17:42
OK.

304
00:17:42 --> 00:17:47
So that's the first goal of my
lecture, was to find this

305
00:17:47 --> 00:17:47
formula.

306
00:17:47 --> 00:17:48
It's done.

307
00:17:48 --> 00:17:48.76
OK.

308
00:17:48.76 --> 00:17:52
And of course I
could invert,

309
00:17:52 --> 00:17:57
now, I can, I sort of like I
can see what -- I can answer

310
00:17:57 --> 00:17:59
questions like this.

311
00:17:59 --> 00:18:04
Suppose I have a matrix,
and let me move the one one

312
00:18:04 --> 00:18:04
entry.

313
00:18:04 --> 00:18:07
What happens to the inverse?

314
00:18:07 --> 00:18:11
Just, just think about that
question.

315
00:18:11 --> 00:18:16
Suppose I have some
matrix, I just write down some

316
00:18:16 --> 00:18:20
nice, non-singular matrix that's
got an inverse,

317
00:18:20 --> 00:18:23
and then I move the one one
entry a little bit.

318
00:18:23 --> 00:18:25
I add one to it,
for example.

319
00:18:25 --> 00:18:28
What happens to the inverse
matrix?

320
00:18:28 --> 00:18:31
Well, this formula should tell
me.

321
00:18:31 --> 00:18:34
I have to look to see what
happens

322
00:18:34 --> 00:18:40
to the determinant and what
happens to all the cofactors.

323
00:18:40 --> 00:18:44
And, the picture,
it's all there.

324
00:18:44 --> 00:18:45
It's all there.

325
00:18:45 --> 00:18:51
We can really understand how
the inverse changes when the

326
00:18:51 --> 00:18:53
matrix changes.

327
00:18:53 --> 00:18:53.37
OK.

328
00:18:53.37 --> 00:18:59
Now my second application is to
-- let me put that over here --

329
00:18:59 --> 00:19:02
is to Ax=b.

330
00:19:02 --> 00:19:07
Well, the -- course,
the solution is A inverse b.

331
00:19:07 --> 00:19:11
But now I have a formula for A
inverse.

332
00:19:11 --> 00:19:16
A inverse is one over the
determinant times this C

333
00:19:16 --> 00:19:18.76
transpose times B.

334
00:19:18.76 --> 00:19:21
I now know what A inverse is.

335
00:19:21 --> 00:19:26
So now I just have to say,
what have I got here?

336
00:19:26 --> 00:19:32
Is there any way to,
to make this formula,

337
00:19:32 --> 00:19:35
this answer,
which is the one and only

338
00:19:35 --> 00:19:41
answer -- it's the very same
answer we got on the first day

339
00:19:41 --> 00:19:43
of the class by elimination.

340
00:19:43 --> 00:19:48
Now I'm -- now I've got a
formula for the answer.

341
00:19:48 --> 00:19:52
Can I play with it further to
see what's going on?

342
00:19:52 --> 00:19:57
And Cramer's,
this Cramer's Rule is

343
00:19:57 --> 00:20:02
exactly, that -- a way of
looking at this formula.

344
00:20:02 --> 00:20:02
OK.

345
00:20:02 --> 00:20:05.08
So this is a formula for x.

346
00:20:05.08 --> 00:20:06
Here's my formula.

347
00:20:06 --> 00:20:08
Well, of course.

348
00:20:08 --> 00:20:14
The first thing I see from the
formula is that the answer x

349
00:20:14 --> 00:20:17
always has that in the
determinant.

350
00:20:17 --> 00:20:19
I'm not surprised.

351
00:20:19 --> 00:20:24
There's a division by the
determinant.

352
00:20:24 --> 00:20:28.22
But then I have to say a little
more carefully what's going on

353
00:20:28.22 --> 00:20:28
up here.

354
00:20:28 --> 00:20:31
And let me tell you what
Cramer's Rule is.

355
00:20:31 --> 00:20:34
Let, let me take x1,
the first component.

356
00:20:34 --> 00:20:36
So this is the first component
of the answer.

357
00:20:36 --> 00:20:40
Then there'll be a second
component and a,

358
00:20:40 --> 00:20:42
all the other components.

359
00:20:42 --> 00:20:47
Can I take just the first
component of this formula?

360
00:20:47 --> 00:20:52
Well, I certainly have
determinant of A down under.

361
00:20:52 --> 00:20:57
And what the heck is the first
-- so what do I get in C

362
00:20:57 --> 00:20:58
transpose b?

363
00:20:58 --> 00:21:02
What's the first entry of C
transpose b?

364
00:21:02 --> 00:21:06
That's what I have to answer
myself.

365
00:21:06 --> 00:21:11.42
Well, what's the first entry of
C transpose b?

366
00:21:11.42 --> 00:21:11
OK.

367
00:21:11 --> 00:21:15
This B is -- let me tell you
what it is.

368
00:21:15 --> 00:21:21.07
Somehow I'm multiplying
cofactors by the entries of B,

369
00:21:21.07 --> 00:21:23
right, in this product.

370
00:21:23 --> 00:21:28
Cofactors from the matrix times
entries of b.

371
00:21:28 --> 00:21:33
So any time I'm multiplying
cofactors by numbers,

372
00:21:33 --> 00:21:37
I think, I'm getting the
determinant of something.

373
00:21:37 --> 00:21:40
And let me call that something
B1.

374
00:21:40 --> 00:21:44
So this is a matrix,
the matrix whose determinant is

375
00:21:44 --> 00:21:46
coming out of that.

376
00:21:46 --> 00:21:49
And we'll,
we'll see what it is.

377
00:21:49 --> 00:21:54
x2 will be the determinant of
some other matrix B2,

378
00:21:54 --> 00:21:57
also divided by determinant of
A.

379
00:21:57 --> 00:22:01
So now I just -- Cramer just
had a good idea.

380
00:22:01 --> 00:22:06
He realized what matrix it was,
what these B1 and B2 and B3 and

381
00:22:06 --> 00:22:08
so on matrices were.

382
00:22:08 --> 00:22:13
Let me write them on the board
underneath.

383
00:22:13 --> 00:22:13
OK.

384
00:22:13 --> 00:22:16
So what is this B1?

385
00:22:16 --> 00:22:24
This B1 is the matrix that has
b in its first column and

386
00:22:24 --> 00:22:28
otherwise the rest of it is A.

387
00:22:28 --> 00:22:37
So it otherwise it has the
rest, the, the n-1 columns of A.

388
00:22:37 --> 00:22:45
It's the matrix with -- it's
just the matrix A with column

389
00:22:45 --> 00:22:51
one replaced by the
right-hand side,

390
00:22:51 --> 00:22:54
by the right-hand side b.

391
00:22:54 --> 00:23:00
Because somehow when I take the
determinant of this guy,

392
00:23:00 --> 00:23:04
it's giving me this matrix
multiplication.

393
00:23:04 --> 00:23:07
Well, how could that be?

394
00:23:07 --> 00:23:13
How could -- so what's,
what's the determinant formula

395
00:23:13 --> 00:23:14.75
I'll use here?

396
00:23:14.75 --> 00:23:19
I'll use cofactors,
of course.

397
00:23:19 --> 00:23:23
And I might as well use
cofactors down column one.

398
00:23:23 --> 00:23:29
So when I use cofactors down
column one, I'm taking the first

399
00:23:29 --> 00:23:31
entry of b times what?

400
00:23:31 --> 00:23:33
Times the cofactor c11.

401
00:23:33 --> 00:23:35
Do you see that?

402
00:23:35 --> 00:23:40
When I, when I use cofactors
here, I take the first entry

403
00:23:40 --> 00:23:43
here,
B one let's call it,

404
00:23:43 --> 00:23:45
times the cofactor there.

405
00:23:45 --> 00:23:50.59
But what's the cofactor in --
my little hand-waving is meant

406
00:23:50.59 --> 00:23:54
to indicate that it's a matrix
of one size smaller,

407
00:23:54 --> 00:23:55
the cofactor.

408
00:23:55 --> 00:23:57
And it's exactly c11.

409
00:23:57 --> 00:23:59
Well, that's just what we
wanted.

410
00:23:59 --> 00:24:03
This first entry is c11
times b1.

411
00:24:03 --> 00:24:08.24
And then the next entry is
whatever, is c21 times b2,

412
00:24:08.24 --> 00:24:09
and so on.

413
00:24:09 --> 00:24:12.03
And sure enough,
if I look here,

414
00:24:12.03 --> 00:24:15
when I'm, when I do the
cofactor expansion,

415
00:24:15 --> 00:24:20
b2 is getting multiplied by the
right thing, and so on.

416
00:24:20 --> 00:24:23
So there's Cramer's Rule.

417
00:24:23 --> 00:24:27.47
And the book gives another kind
of

418
00:24:27.47 --> 00:24:34
cute proof without,
without building so much on,

419
00:24:34 --> 00:24:36
on cofactors.

420
00:24:36 --> 00:24:44
But here we've got cofactors,
so I thought I'd just give you

421
00:24:44 --> 00:24:46
this proof.

422
00:24:46 --> 00:24:52
So what is B -- in general,
what is Bj?

423
00:24:52 --> 00:24:57
This is the,
this is A with column j

424
00:24:57 --> 00:25:01
replaced by, by b.

425
00:25:01 --> 00:25:07
So that's -- the determinant of
that matrix that you divide by

426
00:25:07 --> 00:25:10
determinant of A to get xj.

427
00:25:10 --> 00:25:15
So x -- let me change this
general formula.

428
00:25:15 --> 00:25:19
xj, the j-th one,
is the determinant of Bj

429
00:25:19 --> 00:25:22
divided by the determinant of A.

430
00:25:22 --> 00:25:26.64
And now we've said what Bj is.

431
00:25:26.64 --> 00:25:29
Well, so Cramer found a rule.

432
00:25:29 --> 00:25:33
And we could ask him,
OK, great, good work,

433
00:25:33 --> 00:25:34
Cramer.

434
00:25:34 --> 00:25:38
But is your rule any good in
practice?

435
00:25:38 --> 00:25:42
So he says, well,
you couldn't ask about a rule

436
00:25:42 --> 00:25:47
in mine, right,
because it's just -- all you

437
00:25:47 --> 00:25:54
have to do is find the
determinant of A and these other

438
00:25:54 --> 00:25:57
determinants,
so I guess -- oh,

439
00:25:57 --> 00:26:02
he just says,
well, all you have to do is

440
00:26:02 --> 00:26:07
find n+1 determinants,
the, the n Bs and the A.

441
00:26:07 --> 00:26:12
And actually,
I remember reading -- there was

442
00:26:12 --> 00:26:20
a book, popular book that,
that kids interested in math

443
00:26:20 --> 00:26:24
read when I was a kid
interested in math called

444
00:26:24 --> 00:26:29
Mathematics for the Million or
something, by a guy named Bell.

445
00:26:29 --> 00:26:32
And it had a little page about
linear algebra.

446
00:26:32 --> 00:26:37
And it said,-- so it explained
elimination in a very

447
00:26:37 --> 00:26:38.51
complicated way.

448
00:26:38.51 --> 00:26:41
I certainly didn't understand
it.

449
00:26:41 --> 00:26:44
And, and it made it,
you know, it

450
00:26:44 --> 00:26:48
sort of said,
well, there is this formula for

451
00:26:48 --> 00:26:52
elimination, but look at this
great formula,

452
00:26:52 --> 00:26:53
Cramer's Rule.

453
00:26:53 --> 00:26:58
So it certainly said Cramer's
Rule was the way to go.

454
00:26:58 --> 00:27:02
But actually,
Cramer's Rule is a disastrous

455
00:27:02 --> 00:27:05
way to go, because to compute
these

456
00:27:05 --> 00:27:08
determinants,
it takes, like,

457
00:27:08 --> 00:27:10
approximately forever.

458
00:27:10 --> 00:27:15
So actually I now think of that
book title as being Mathematics

459
00:27:15 --> 00:27:19
for the Millionaire,
because you'd have to be able

460
00:27:19 --> 00:27:24
to pay for, a hopelessly long
calculation

461
00:27:24 --> 00:27:27
where elimination,
of course, produced the x-s,

462
00:27:27 --> 00:27:28
in an instant.

463
00:27:28 --> 00:27:33
But having a formula allows you
to, with, with letters,

464
00:27:33 --> 00:27:36
you know, allows you to do
algebra instead of,

465
00:27:36 --> 00:27:37
algorithms.

466
00:27:37 --> 00:27:42
So the, there's some value in
the Cramer's Rule formula for x

467
00:27:42 --> 00:27:47
and in the explicit formula
for, for A inverse.

468
00:27:47 --> 00:27:52.42
They're nice formulas,
but I just don't want you to

469
00:27:52.42 --> 00:27:53.39
use them.

470
00:27:53.39 --> 00:27:56
That'ss what it comes to.

471
00:27:56 --> 00:28:01
If you had to -- and Matlab
would never, never do it.

472
00:28:01 --> 00:28:06
I mean, it would use
elimination.

473
00:28:06 --> 00:28:06
OK.

474
00:28:06 --> 00:28:11
Now I'm ready for number three
in today's list of amazing

475
00:28:11 --> 00:28:16
connections coming through the
determinant.

476
00:28:16 --> 00:28:22
And that number three is the
fact that the determinant gives

477
00:28:22 --> 00:28:22
a volume.

478
00:28:22 --> 00:28:23
OK.

479
00:28:23 --> 00:28:29
So now -- so that's my final
topic for -- among these --

480
00:28:29 --> 00:28:33
this my number three
application, that the

481
00:28:33 --> 00:28:39
determinant is actually equals
the volume of something.

482
00:28:39 --> 00:28:44
Can I use this little space to
consider a special case,

483
00:28:44 --> 00:28:50
and then I'll use the far board
to think about the general rule.

484
00:28:50 --> 00:28:53
So what I going to prove?

485
00:28:53 --> 00:28:54
Or claim.

486
00:28:54 --> 00:29:00
I claim that the determinant of
the matrix is the volume of a

487
00:29:00 --> 00:29:01
box.

488
00:29:01 --> 00:29:03
OK, and you say,
which box?

489
00:29:03 --> 00:29:04
Fair enough.

490
00:29:04 --> 00:29:05
OK.

491
00:29:05 --> 00:29:06
So let's see.

492
00:29:06 --> 00:29:11
I'm in -- shall we say we're
in, say three by three?

493
00:29:11 --> 00:29:18
Shall we suppose -- let's,
let's say three by three.

494
00:29:18 --> 00:29:23
So, so we can really -- we're,
we're talking about boxes in

495
00:29:23 --> 00:29:27
three dimensions,
and three by three matrices.

496
00:29:27 --> 00:29:32
And so all I do -- you could
guess what the box is.

497
00:29:32 --> 00:29:35
Here is, here is,
three dimensions.

498
00:29:35 --> 00:29:35
OK.

499
00:29:35 --> 00:29:41
Now I take the first row of the
matrix, a11, a22,

500
00:29:41 --> 00:29:43
A -- sorry.
a11, a12, a13.

501
00:29:43 --> 00:29:44
That row is a vector.

502
00:29:44 --> 00:29:46
It goes to some point.

503
00:29:46 --> 00:29:50
That point will be -- and that
edge going to it,

504
00:29:50 --> 00:29:55
will be an edge of the box,
and that point will be a corner

505
00:29:55 --> 00:29:56
of the box.

506
00:29:56 --> 00:30:00
So here is zero zero zero,
of course.

507
00:30:00 --> 00:30:05
And here's the first row of the
matrix: a11, a12,

508
00:30:05 --> 00:30:06
a13.

509
00:30:06 --> 00:30:09
So that's one edge of the box.

510
00:30:09 --> 00:30:16
Another edge of the box is to
the second row of the matrix,

511
00:30:16 --> 00:30:17
row two.

512
00:30:17 --> 00:30:20.93
Can I just call it there row
two?

513
00:30:20.93 --> 00:30:28
And a third row of the box will
be to -- a third row --

514
00:30:28 --> 00:30:33
a third edge of the box will be
given by row three.

515
00:30:33 --> 00:30:35
So, so there's row three.

516
00:30:35 --> 00:30:40
That, the coordinates,
what are the coordinates of

517
00:30:40 --> 00:30:43
that corner of the box?
a31, a32, a33.

518
00:30:43 --> 00:30:48
So I've got that edge of the
box, that edge of the box,

519
00:30:48 --> 00:30:54
that edge of
the box, and that's all I need.

520
00:30:54 --> 00:30:58
Now I just finish out the box,
right?

521
00:30:58 --> 00:31:04
I just -- the proper word,
of course, is parallelepiped.

522
00:31:04 --> 00:31:08
But for obvious reasons,
I wrote box.

523
00:31:08 --> 00:31:09
OK.

524
00:31:09 --> 00:31:10
So, OK.

525
00:31:10 --> 00:31:16
So there's the,
there's the bottom of the box.

526
00:31:16 --> 00:31:20
There're the four edge sides of
the box.

527
00:31:20 --> 00:31:23
There's the top of the box.

528
00:31:23 --> 00:31:24
Cute, right?

529
00:31:24 --> 00:31:29
It's the box that has these
three edges and then it's

530
00:31:29 --> 00:31:33
completed to a,
to a, each, you know,

531
00:31:33 --> 00:31:37
each side is a,
is a parallelogram.

532
00:31:37 --> 00:31:43
And it's that box whose volume
is given by the determinant.

533
00:31:43 --> 00:31:49
That's -- now it's --
possible that the determinant

534
00:31:49 --> 00:31:50
is negative.

535
00:31:50 --> 00:31:55
So we have to just say what's
going on in that case.

536
00:31:55 --> 00:31:59
If the determinant is negative,
then the volume,

537
00:31:59 --> 00:32:03
we, we should take the absolute
value really.

538
00:32:03 --> 00:32:08
So the volume,
if we, if we think of volume as

539
00:32:08 --> 00:32:11
positive, we should take the
absolute value of the

540
00:32:11 --> 00:32:12
determinant.

541
00:32:12 --> 00:32:14
But the, the sign,
what does the sign of the

542
00:32:14 --> 00:32:17
determinant -- it always must
tell us something.

543
00:32:17 --> 00:32:20
And somehow it,
it will tell us whether these

544
00:32:20 --> 00:32:23
three is a -- whether it's a
right-handed box or a

545
00:32:23 --> 00:32:24
left-handed box.

546
00:32:24 --> 00:32:27
If we, if we reversed two of
the

547
00:32:27 --> 00:32:31
edges, we would go between a
right-handed box and a

548
00:32:31 --> 00:32:33
left-handed box.

549
00:32:33 --> 00:32:38
We wouldn't change the volume,
but we would change the,

550
00:32:38 --> 00:32:39
the cyclic, order.

551
00:32:39 --> 00:32:42
So I won't worry about that.

552
00:32:42 --> 00:32:45
And, so one special case is
what?

553
00:32:45 --> 00:32:47
A equal identity matrix.

554
00:32:47 --> 00:32:50.09
So let's take that special
case.

555
00:32:50.09 --> 00:32:52
A equal identity matrix.

556
00:32:52 --> 00:32:57
Is the
formula determinant of identity

557
00:32:57 --> 00:33:01
matrix, does that equal the
volume of the box?

558
00:33:01 --> 00:33:04
Well, what is the box?

559
00:33:04 --> 00:33:05
What's the box?

560
00:33:05 --> 00:33:11
If A is the identity matrix,
then these three rows are the

561
00:33:11 --> 00:33:17.25
three coordinate vectors,
and the box is -- it's a cube.

562
00:33:17.25 --> 00:33:20
It's the unit cube.

563
00:33:20 --> 00:33:24.4
So if, if A is the identity
matrix, of course our formula is

564
00:33:24.4 --> 00:33:24
right.

565
00:33:24 --> 00:33:28
Well, actually that proves
property one -- that the volume

566
00:33:28 --> 00:33:30
has property one.

567
00:33:30 --> 00:33:33
Actually, we could,
we could, we could get this

568
00:33:33 --> 00:33:37
thing if we -- if we can show
that the box volume has the same

569
00:33:37 --> 00:33:41
three properties that define the
determinant,

570
00:33:41 --> 00:33:43
then it must be the
determinant.

571
00:33:43 --> 00:33:47
So that's like the,
the, the elegant way to prove

572
00:33:47 --> 00:33:47.91
this.

573
00:33:47.91 --> 00:33:52
To prove this amazing fact that
the determinant equals the

574
00:33:52 --> 00:33:56
volume, first we'll check it for
the identity matrix.

575
00:33:56 --> 00:33:57
That's fine.

576
00:33:57 --> 00:34:00
The box is a cube and its
volume is one and the

577
00:34:00 --> 00:34:04
determinant is one and,
and one agrees with one.

578
00:34:04 --> 00:34:08
Now let me
take one -- let me go up one

579
00:34:08 --> 00:34:10
level to an orthogonal matrix.

580
00:34:10 --> 00:34:15
Because I'd like to take this
chance to bring in chapter --

581
00:34:15 --> 00:34:18.06
the, the previous chapter.

582
00:34:18.06 --> 00:34:21
Suppose I have an orthogonal
matrix.

583
00:34:21 --> 00:34:22
What did that mean?

584
00:34:22 --> 00:34:25.26
I always called those things Q.

585
00:34:25.26 --> 00:34:28
What was the point of --
suppose I

586
00:34:28 --> 00:34:35
have, suppose instead of the
identity matrix I'm now going to

587
00:34:35 --> 00:34:39
take A equal Q,
an orthogonal matrix.

588
00:34:39 --> 00:34:41
What was Q then?

589
00:34:41 --> 00:34:47
That was a matrix whose columns
were orthonormal,

590
00:34:47 --> 00:34:47.69
right?

591
00:34:47.69 --> 00:34:52
Those were its columns were
unit vectors,

592
00:34:52 --> 00:34:56
perpendicular unit vectors.

593
00:34:56 --> 00:34:59
So what kind of a box have we
got now?

594
00:34:59 --> 00:35:04
What kind of a box comes from
the rows or the columns,

595
00:35:04 --> 00:35:07
I don't mind,
because the determinant is the

596
00:35:07 --> 00:35:12
determinant of the transpose,
so I'm never worried about

597
00:35:12 --> 00:35:12
that.

598
00:35:12 --> 00:35:16.84
What kind of a box,
what shape box have we got if

599
00:35:16.84 --> 00:35:20
the matrix is an
orthogonal matrix?

600
00:35:20 --> 00:35:22
It's another cube.

601
00:35:22 --> 00:35:24
It's a cube again.

602
00:35:24 --> 00:35:28.41
How is it different from the
identity cube?

603
00:35:28.41 --> 00:35:30
It's just rotated.

604
00:35:30 --> 00:35:35
It's just the orthogonal matrix
Q doesn't have to be the

605
00:35:35 --> 00:35:36
identity matrix.

606
00:35:36 --> 00:35:41
It's just the unit cube but
turned in space.

607
00:35:41 --> 00:35:45
So sure enough,
it's the unit cube,

608
00:35:45 --> 00:35:47.41
and its volume is one.

609
00:35:47.41 --> 00:35:49
Now is the determinant one?

610
00:35:49 --> 00:35:52
What's the determinant of Q?

611
00:35:52 --> 00:35:57
We believe that the determinant
of Q better be one or minus one,

612
00:35:57 --> 00:36:01
so that our formula is --
checks out in that -- if we

613
00:36:01 --> 00:36:06
can't check it in these easy
cases where we got a cube,

614
00:36:06 --> 00:36:09
we're not going to get it in
the

615
00:36:09 --> 00:36:10
general case.

616
00:36:10 --> 00:36:15
So why is the determinant of Q
equal one or minus one?

617
00:36:15 --> 00:36:17
What do we know about Q?

618
00:36:17 --> 00:36:22
What's the one matrix statement
of the properties of Q?

619
00:36:22 --> 00:36:27
A matrix with orthonormal
columns has -- satisfies a

620
00:36:27 --> 00:36:29
certain equation.

621
00:36:29 --> 00:36:30
What, what is that?

622
00:36:30 --> 00:36:35
It's if we
have this orthogonal matrix,

623
00:36:35 --> 00:36:42
then the fact -- the way to say
what it, what its properties are

624
00:36:42 --> 00:36:43
is this.

625
00:36:43 --> 00:36:47
Q prime, u- u- Q transpose Q
equals I.

626
00:36:47 --> 00:36:48
Right?

627
00:36:48 --> 00:36:54
That's what -- those are the
matrices that get the name Q,

628
00:36:54 --> 00:36:59
the matrices that
Q transpose Q is I.

629
00:36:59 --> 00:37:00
OK.

630
00:37:00 --> 00:37:04
Now from that,
tell me why is the determinant

631
00:37:04 --> 00:37:06
one or minus one.

632
00:37:06 --> 00:37:11
How do I, out of this fact --
this may even be a homework

633
00:37:11 --> 00:37:12
problem.

634
00:37:12 --> 00:37:17
It's there in the,
in the list of exercises in the

635
00:37:17 --> 00:37:20
book, and let's just do it.

636
00:37:20 --> 00:37:24
How do I get,
how do I discover

637
00:37:24 --> 00:37:29
that the determinant of Q is
one or maybe minus one?

638
00:37:29 --> 00:37:33
I take determinants of both
sides, everybody says,

639
00:37:33 --> 00:37:37
so I won't -- I take
determinants of both sides.

640
00:37:37 --> 00:37:41
On the right-hand side -- so I,
when I take determinants of

641
00:37:41 --> 00:37:44
both sides, let me just do it.

642
00:37:44 --> 00:37:49
Take the determinant of --
take determinants.

643
00:37:49 --> 00:37:52
Determinant of the identity is
one.

644
00:37:52 --> 00:37:55
What's the determinant of that
product?

645
00:37:55 --> 00:37:58
Rule nine is paying off now.

646
00:37:58 --> 00:38:04
The determinant of a product is
the determinant of this guy --

647
00:38:04 --> 00:38:08
maybe I'll put it,
I'll use that symbol for

648
00:38:08 --> 00:38:09
determinant.

649
00:38:09 --> 00:38:13
It's the determinant of that
guy

650
00:38:13 --> 00:38:16
times the determinant of the
other guy.

651
00:38:16 --> 00:38:21
And then what's the determinant
of Q transpose?

652
00:38:21 --> 00:38:24
It's the same as the
determinant of Q.

653
00:38:24 --> 00:38:26
Rule ten pays off.

654
00:38:26 --> 00:38:29
So this is just this thing
squared.

655
00:38:29 --> 00:38:35
So that determinant squared is
one and sure enough it's one or

656
00:38:35 --> 00:38:36
minus one.

657
00:38:36 --> 00:38:36
Great.

658
00:38:36 --> 00:38:41
So in these special cases of
cubes,

659
00:38:41 --> 00:38:46
we really do have determinant
equals volume.

660
00:38:46 --> 00:38:50
Now can I just push that to
non-cubes.

661
00:38:50 --> 00:38:56
Let me push it first to
rectangles, rectangular boxes,

662
00:38:56 --> 00:39:03
where I'm just multiplying the
e- the edges are -- let me keep

663
00:39:03 --> 00:39:10
all the ninety degree angles,
because those are -- that,

664
00:39:10 --> 00:39:14
that makes my life easy.

665
00:39:14 --> 00:39:16
And just stretch the edges.

666
00:39:16 --> 00:39:21
Suppose I stretch that first
edge, suppose this first edge I

667
00:39:21 --> 00:39:22
double.

668
00:39:22 --> 00:39:26
Suppose I double that first
edge, keeping the other edges

669
00:39:26 --> 00:39:27
the same.

670
00:39:27 --> 00:39:29
What happens to the volume?

671
00:39:29 --> 00:39:31
It doubles, right?

672
00:39:31 --> 00:39:34
We know that the volume of a
cube doubles.

673
00:39:34 --> 00:39:38
In fact, because we know that
the

674
00:39:38 --> 00:39:42
new cube would sit right on top
-- I mean, the new,

675
00:39:42 --> 00:39:46
the added cube would sit right
on -- would fit -- probably a

676
00:39:46 --> 00:39:51
geometer would say congruent or
something -- would go right in,

677
00:39:51 --> 00:39:52
in the other.

678
00:39:52 --> 00:39:53
We'd have two.

679
00:39:53 --> 00:39:55
We have two identical cubes.

680
00:39:55 --> 00:39:57
Total volume is now two.

681
00:39:57 --> 00:39:58
OK.

682
00:39:58 --> 00:40:04
So I want -- if I double an
edge, the volume doubles.

683
00:40:04 --> 00:40:07
What happens to the
determinant?

684
00:40:07 --> 00:40:11.75
If I double,
the first row of a matrix,

685
00:40:11.75 --> 00:40:17
what ch- ch- what's the effect
on the determinant?

686
00:40:17 --> 00:40:19
It also doubles,
right?

687
00:40:19 --> 00:40:22
And that was rule number 3a.

688
00:40:22 --> 00:40:30
Remember rule 3a was that if I,
I could, if I had a factor in,

689
00:40:30 --> 00:40:34
in row one, T,
I could factor it out.

690
00:40:34 --> 00:40:38
So if, if I have a factor two
in that row one,

691
00:40:38 --> 00:40:42
I can factor it out of the
determinant.

692
00:40:42 --> 00:40:47
It agrees with the -- the
volume of the box has that

693
00:40:47 --> 00:40:48
factor two.

694
00:40:48 --> 00:40:53
So, so volume satisfies this
property 3a.

695
00:40:53 --> 00:40:59
And now I really close,
but I -- but to get to the very

696
00:40:59 --> 00:41:04
end of this proof,
I have to get away from right

697
00:41:04 --> 00:41:04
angles.

698
00:41:04 --> 00:41:09
I have to allow the possibility
of, other angles.

699
00:41:09 --> 00:41:15.71
And -- or what's saying the
same thing, I have to check that

700
00:41:15.71 --> 00:41:18
the volume also satisfies 3b.

701
00:41:18 --> 00:41:24
So can I --
This is end of proof that the

702
00:41:24 --> 00:41:29
-- so I'm -- determinant of A
equals volume of box,

703
00:41:29 --> 00:41:32
and where I right now?

704
00:41:32 --> 00:41:37.25
This volume has properties,
properties one,

705
00:41:37.25 --> 00:41:38
no problem.

706
00:41:38 --> 00:41:44
If the box is the cube,
everything is -- if the box is

707
00:41:44 --> 00:41:49
the unit cube,
its volume is one.

708
00:41:49 --> 00:41:53
Property two was if I reverse
two rows, but that doesn't

709
00:41:53 --> 00:41:55
change the box.

710
00:41:55 --> 00:41:58
And it doesn't change the
absolute value,

711
00:41:58 --> 00:41:59
so no problem there.

712
00:41:59 --> 00:42:03
Property 3a was if I mul- you
remember what 3a was?

713
00:42:03 --> 00:42:07.53
So property one was about the
identity matrix.

714
00:42:07.53 --> 00:42:12.34
Property two was about a plus
or minus sign that I don't care

715
00:42:12.34 --> 00:42:13
about.

716
00:42:13 --> 00:42:17
Property 3a was a factor T in a
row.

717
00:42:17 --> 00:42:22
But now I've got property three
B to deal with.

718
00:42:22 --> 00:42:24
What was property 3b?

719
00:42:24 --> 00:42:28
This is a great way to review
these, properties.

720
00:42:28 --> 00:42:33
So that 3b, the property 3b
said -- let's do,

721
00:42:33 --> 00:42:36
let's do two by two.

722
00:42:36 --> 00:42:43
So said that if I had a+a',
b+b', c, d that this equaled

723
00:42:43 --> 00:42:44.13
what?

724
00:42:44.13 --> 00:42:47
So this is property 3b.

725
00:42:47 --> 00:42:52
This is the linearity in row
one by itself.

726
00:42:52 --> 00:42:59
So c d is staying the same,
and I can split this into a b

727
00:42:59 --> 00:43:01
and a' b'.

728
00:43:01 --> 00:43:06
That's property 3b,
at least in the two by two

729
00:43:06 --> 00:43:09
case.

730
00:43:09 --> 00:43:14
And what I -- I wanted now to
show that the volume,

731
00:43:14 --> 00:43:19
which two, two by two,
that means area,

732
00:43:19 --> 00:43:22
has this, has this property.

733
00:43:22 --> 00:43:28
Let me just emphasize that we
have got -- we're getting --

734
00:43:28 --> 00:43:34
this is a formula,
then, for the area of a

735
00:43:34 --> 00:43:36
parallelogram.

736
00:43:36 --> 00:43:41
The area of this parallelogram
-- can I just draw it?

737
00:43:41 --> 00:43:45
OK, here's the,
here's the parallelogram.

738
00:43:45 --> 00:43:47
I have the row a b.

739
00:43:47 --> 00:43:49
That's the first row.

740
00:43:49 --> 00:43:52
That's the point a b.

741
00:43:52 --> 00:43:56
And I tack on c d.
c d, coming out of here.

742
00:43:56 --> 00:44:01
And I complete the
parallelogram.

743
00:44:01 --> 00:44:06
So this is -- well,
I better make it look right.

744
00:44:06 --> 00:44:12.71
It's really this one that has
coordinates c d and this has

745
00:44:12.71 --> 00:44:17
coordinates -- well,
whatever the sum is.

746
00:44:17 --> 00:44:21
And of course starting at zero
zero.

747
00:44:21 --> 00:44:24
So we all know,
this is a+c,

748
00:44:24 --> 00:44:25
b+d.

749
00:44:25 --> 00:44:31
Rather than -- I'm pausing on
that proof for a minute just to

750
00:44:31 --> 00:44:33
going back to our formula.

751
00:44:33 --> 00:44:38
Because I want you to see that
unlike Cramer's Rule,

752
00:44:38 --> 00:44:44.29
that I wasn't that impressed
by, I'm very impressed by this

753
00:44:44.29 --> 00:44:47
formula for the area of a
parallelogram.

754
00:44:47 --> 00:44:50
And what's our formula?

755
00:44:50 --> 00:44:53.46
What,
what's the area of that

756
00:44:53.46 --> 00:44:54
parallelogram?

757
00:44:54 --> 00:44:58
If I had asked you that last
year, you would have said OK,

758
00:44:58 --> 00:45:03
the area of a parallelogram is
the base times the height,

759
00:45:03 --> 00:45:03
right?

760
00:45:03 --> 00:45:06
So you would have figured out
what this base,

761
00:45:06 --> 00:45:09.24
the -- how long that base was.

762
00:45:09.24 --> 00:45:14
It's like the square root of A
squared plus b squared.

763
00:45:14 --> 00:45:21
And then you would have figured
out how much is this height,

764
00:45:21 --> 00:45:22
whatever it is.

765
00:45:22 --> 00:45:24
It's horrible.

766
00:45:24 --> 00:45:27
This, I mean,
we got square roots,

767
00:45:27 --> 00:45:33.7
and in that height there would
be other revolting stuff.

768
00:45:33.7 --> 00:45:40
But now what's the formula that
we now know for the area?

769
00:45:40 --> 00:45:46
It's the determinant of our
little matrix.

770
00:45:46 --> 00:45:48
It's just ad-bc.

771
00:45:48 --> 00:45:50
No square roots.

772
00:45:50 --> 00:45:57
Totally rememberable,
because it's exactly a formula

773
00:45:57 --> 00:46:04
that we've been studying the
whole, for three lectures.

774
00:46:04 --> 00:46:05
OK.

775
00:46:05 --> 00:46:11
That's, you know,
that's the most important point

776
00:46:11 --> 00:46:15
I'm making here.

777
00:46:15 --> 00:46:19.31
Is that if you know the
coordinates of a box,

778
00:46:19.31 --> 00:46:23
of the corners,
then you have a great formula

779
00:46:23 --> 00:46:25
for the volume,
area or volume,

780
00:46:25 --> 00:46:30
that doesn't involve any
lengths or any angles or any

781
00:46:30 --> 00:46:36
heights, but just involves the
coordinates that you've got.

782
00:46:36 --> 00:46:38
And similarly,
what's the area of this

783
00:46:38 --> 00:46:39
triangle?

784
00:46:39 --> 00:46:44
Suppose I chop that off and say
what about -- because you might

785
00:46:44 --> 00:46:47
often be interested in a
triangle instead of a

786
00:46:47 --> 00:46:48
parallelogram.

787
00:46:48 --> 00:46:51
What's the area of this
triangle?

788
00:46:51 --> 00:46:54
Now there again,
everybody would have said the

789
00:46:54 --> 00:46:59
area of a triangle is
half the base times the height.

790
00:46:59 --> 00:47:03
And in some cases,
if you know the base that a,

791
00:47:03 --> 00:47:06
that's -- and the height,
that's fine.

792
00:47:06 --> 00:47:09
But here, we,
what we know is the coordinates

793
00:47:09 --> 00:47:11
of the corners.

794
00:47:11 --> 00:47:12
We know the vertices.

795
00:47:12 --> 00:47:15
And so what's the area of that
triangle?

796
00:47:15 --> 00:47:19
If I know these,
if I know a b,

797
00:47:19 --> 00:47:24
c d, and zero zero,
what's the area?

798
00:47:24 --> 00:47:28
It's just half,
so it's just half of this.

799
00:47:28 --> 00:47:36
So this is, this is a- a b -- a
d - b c for the parallelogram

800
00:47:36 --> 00:47:41
and one half of that,
one half of ad-bc for the

801
00:47:41 --> 00:47:42
triangle.

802
00:47:42 --> 00:47:47
So I mean,
this is a totally trivial

803
00:47:47 --> 00:47:50
remark, to say,
well, divide by two.

804
00:47:50 --> 00:47:53
But it's just that you more
often see triangles,

805
00:47:53 --> 00:47:58
and you feel you know the
formula for the area but the

806
00:47:58 --> 00:48:01
good formula for the area is
this one.

807
00:48:01 --> 00:48:05.97
And I'm just going to -- I'm
just going to say one more thing

808
00:48:05.97 --> 00:48:09
about the area of a triangle.

809
00:48:09 --> 00:48:15
It's just because it's -- you
know, it's so great to have a

810
00:48:15 --> 00:48:18
good formula for something.

811
00:48:18 --> 00:48:22
What if our triangle did not
start at zero zero?

812
00:48:22 --> 00:48:28
What if our triangle,
what if we had this -- what if

813
00:48:28 --> 00:48:33
we had -- so I'm coming back to
triangles again.

814
00:48:33 --> 00:48:38
But let me, let me put this
triangle somewhere,

815
00:48:38 --> 00:48:44
it's -- I'm staying with
triangles, I'm just in two

816
00:48:44 --> 00:48:50
dimensions, but I'm going to
allow you to give me any three

817
00:48:50 --> 00:48:51
corners.

818
00:48:51 --> 00:48:57
And in -- those six numbers
must determine the area.

819
00:48:57 --> 00:48:59
And what's the formula?

820
00:48:59 --> 00:49:04
The area is going to be,
it's going to be,

821
00:49:04 --> 00:49:09
there'll be that
half of a parallelogram.

822
00:49:09 --> 00:49:13.68
I mean, basically this can't be
completely new,

823
00:49:13.68 --> 00:49:14
right?

824
00:49:14 --> 00:49:19
We've got the area when -- we,
we know the area when this is

825
00:49:19 --> 00:49:20
zero zero.

826
00:49:20 --> 00:49:25
Now we just want to lift our
sight slightly and get the area

827
00:49:25 --> 00:49:29
when all th- so let me write
down what it,

828
00:49:29 --> 00:49:32
what it comes out to be.

829
00:49:32 --> 00:49:36
It turns out that if you do
this, x1 y1 and a 1,

830
00:49:36 --> 00:49:39
x2 y2 and a 1,
x3 y3 and a 1,

831
00:49:39 --> 00:49:40.73
that that works.

832
00:49:40.73 --> 00:49:44
That the determinant symbol,
of course.

833
00:49:44 --> 00:49:49.04
It's just -- if I gave you that
determinant to find,

834
00:49:49.04 --> 00:49:52
you might subtract this row
from

835
00:49:52 --> 00:49:53
this.

836
00:49:53 --> 00:49:55
It would kill that one.

837
00:49:55 --> 00:49:59
Subtract this row from this,
it would kill that one.

838
00:49:59 --> 00:50:03.46
Then you'd have a simple
determinant to do with

839
00:50:03.46 --> 00:50:07
differences, and it would --
this little subtraction,

840
00:50:07 --> 00:50:12
what I did was equivalent to
moving the triangle to start at

841
00:50:12 --> 00:50:13
the origin.

842
00:50:13 --> 00:50:17
I did it fast,
because time is up.

843
00:50:17 --> 00:50:21
And I didn't complete that
proof of 3b.

844
00:50:21 --> 00:50:27
I'll leave -- the book has a
carefully drawn figure to show

845
00:50:27 --> 00:50:29
why that works.

846
00:50:29 --> 00:50:34
But I hope you saw the main
point is that for area and

847
00:50:34 --> 00:50:38
volume, determinant gives a
great formula.

848
00:50:38 --> 00:50:38
OK.

849
00:50:38 --> 00:50:42
And
next lectures are about

850
00:50:42 --> 00:50:47
eigenvalues, so we're really
into the big stuff.

851
00:50:47 --> 00:50:50
Thanks.