1
00:00:00,530 --> 00:00:02,960
The following content is
provided under a Creative

2
00:00:02,960 --> 00:00:04,370
Commons license.

3
00:00:04,370 --> 00:00:07,410
Your support will help MIT
OpenCourseWare continue to

4
00:00:07,410 --> 00:00:11,060
offer high quality educational
resources for free.

5
00:00:11,060 --> 00:00:13,960
To make a donation or view
additional materials from

6
00:00:13,960 --> 00:00:16,680
hundreds of MIT courses, visit
mitopencourseware@ocw.mit.edu.

7
00:00:34,810 --> 00:00:36,870
PROFESSOR: So last week
Melissa told you--

8
00:00:36,870 --> 00:00:38,140
or a couple days ago--

9
00:00:38,140 --> 00:00:39,510
about vision.

10
00:00:39,510 --> 00:00:42,780
And if there's two huge
messages, I think, out of that

11
00:00:42,780 --> 00:00:46,210
lecture, it's first that even
the most simple aspects of

12
00:00:46,210 --> 00:00:50,150
perception are based on a
synthesis between what's out

13
00:00:50,150 --> 00:00:53,000
there and the rules
by which your mind

14
00:00:53,000 --> 00:00:55,130
interprets the world.

15
00:00:55,130 --> 00:00:58,510
And what you ultimately see
is that interaction, that

16
00:00:58,510 --> 00:01:02,010
synthesis, that dialogue, that
construction between what's

17
00:01:02,010 --> 00:01:05,620
out there and how your mind
chooses to interpret it.

18
00:01:05,620 --> 00:01:07,990
The second big point is there's
a lot of organization

19
00:01:07,990 --> 00:01:09,890
of the visual system
in our brain.

20
00:01:09,890 --> 00:01:13,240
And we can track its
organization from the first

21
00:01:13,240 --> 00:01:17,460
moment that photons enter your
eyes all the way through until

22
00:01:17,460 --> 00:01:22,940
you recognize a loved one, a
word, or an object out there.

23
00:01:22,940 --> 00:01:27,040
And we said that in the
neocortex, the world is

24
00:01:27,040 --> 00:01:31,190
divided into two big channels
of visual processing, which

25
00:01:31,190 --> 00:01:34,990
are what things are and
where things are.

26
00:01:34,990 --> 00:01:38,220
So now I'm going to pick up on
that a little bit today.

27
00:01:38,220 --> 00:01:41,370
As you go around and see things
all the time, what is

28
00:01:41,370 --> 00:01:43,850
it that you mostly see?

29
00:01:43,850 --> 00:01:46,430
On an average day, you wake
up in the morning.

30
00:01:50,795 --> 00:01:54,910
I know you're all like,
what's he looking for?

31
00:01:54,910 --> 00:01:56,140
What's the answer?

32
00:01:56,140 --> 00:01:57,260
But just in everything.

33
00:01:57,260 --> 00:02:00,740
You see people a lot.

34
00:02:00,740 --> 00:02:02,070
So faces, we'll talk
about that.

35
00:02:02,070 --> 00:02:04,790
You see objects, chairs,
tables in the morning.

36
00:02:04,790 --> 00:02:06,690
A toothbrush and soap
is useful, right?

37
00:02:06,690 --> 00:02:07,980
You see objects.

38
00:02:07,980 --> 00:02:11,070
And another huge use of your
visual system is reading.

39
00:02:11,070 --> 00:02:14,040
You see words, right, on
books, on computers.

40
00:02:14,040 --> 00:02:18,050
So we're going to talk a little
bit about how we see

41
00:02:18,050 --> 00:02:21,310
what things are, objects in the
world, faces in the world,

42
00:02:21,310 --> 00:02:23,790
and words so that we can read.

43
00:02:23,790 --> 00:02:26,480
And we'll talk about both some
processes that go on in our

44
00:02:26,480 --> 00:02:29,610
mind and brain that let us
become terrific at recognizing

45
00:02:29,610 --> 00:02:33,080
what's out there and knowing
who we're talking to, what

46
00:02:33,080 --> 00:02:36,110
we're reading, what we're
holding, and so on, and then

47
00:02:36,110 --> 00:02:38,540
about disorders in the brain
that illuminate, I think, a

48
00:02:38,540 --> 00:02:40,990
little bit about those
processes.

49
00:02:40,990 --> 00:02:46,970
So let me start with objects
and agnosia.

50
00:02:46,970 --> 00:02:50,110
Agnosia means not knowing.

51
00:02:50,110 --> 00:02:52,680
So we're going to talk about two
ways in which people seen

52
00:02:52,680 --> 00:02:53,440
to not know.

53
00:02:53,440 --> 00:02:54,910
And that gives us a hint,
then, about how

54
00:02:54,910 --> 00:02:56,980
we actually do know.

55
00:02:56,980 --> 00:02:58,630
And they've been called
apperceptive

56
00:02:58,630 --> 00:02:59,990
and associative agnosias.

57
00:02:59,990 --> 00:03:02,400
Apperceptive seems to be people
see the parts, but they

58
00:03:02,400 --> 00:03:03,640
don't pull it together.

59
00:03:03,640 --> 00:03:05,500
Associative agnosia,
they see the whole

60
00:03:05,500 --> 00:03:06,830
thing pretty decently.

61
00:03:06,830 --> 00:03:10,390
But they can't figure out what
it means, what it is in a

62
00:03:10,390 --> 00:03:11,640
practical sense.

63
00:03:11,640 --> 00:03:13,880
And, lastly, I'll talk to you
about a very surprising

64
00:03:13,880 --> 00:03:17,070
finding where people lose
knowledge of some kinds of

65
00:03:17,070 --> 00:03:20,180
objects in the world but not
others, and sort of give us a

66
00:03:20,180 --> 00:03:24,170
new insight into how our brain
organizes knowledge.

67
00:03:24,170 --> 00:03:26,400
So agnosia is a
modality-specific inability.

68
00:03:26,400 --> 00:03:27,840
So we'll talk today
about vision.

69
00:03:27,840 --> 00:03:32,050
There's agnosias in audition,
and in touch, and so on.

70
00:03:32,050 --> 00:03:34,210
But I'll focus just on vision,
modality-specific.

71
00:03:34,210 --> 00:03:35,880
These people are fine
if they hold

72
00:03:35,880 --> 00:03:37,730
something or hear something.

73
00:03:37,730 --> 00:03:39,520
And it's not explained by things
like you don't see so

74
00:03:39,520 --> 00:03:42,600
well, or you're not paying
attention, or you can't speak.

75
00:03:42,600 --> 00:03:45,640
It seems to be just that you
lose the ability to recognize

76
00:03:45,640 --> 00:03:48,960
by sight what something is.

77
00:03:48,960 --> 00:03:51,510
So again, apperceptive agnosia,
will link it to the

78
00:03:51,510 --> 00:03:54,460
right hemisphere and associative
agnosia to the

79
00:03:54,460 --> 00:03:55,555
left hemisphere.

80
00:03:55,555 --> 00:03:58,400
And let me show you examples
of what these patients do.

81
00:03:58,400 --> 00:04:00,620
So here's patients with right
hemisphere damage.

82
00:04:00,620 --> 00:04:04,430
We're attempting, simply, to
copy this triangle or copy

83
00:04:04,430 --> 00:04:04,930
these things.

84
00:04:04,930 --> 00:04:05,760
It's right in front of them.

85
00:04:05,760 --> 00:04:06,830
They're copying it.

86
00:04:06,830 --> 00:04:09,040
And you can see that it's as
if they have almost no

87
00:04:09,040 --> 00:04:10,660
appreciation for what the
thing is out there.

88
00:04:10,660 --> 00:04:13,620
This is a pretty miserable 3.

89
00:04:13,620 --> 00:04:16,740
Right in front of you, it's as
if the whole thing weren't

90
00:04:16,740 --> 00:04:17,899
cohering for anything.

91
00:04:17,899 --> 00:04:22,680
Even a shape is just a
mysterious bundle of stuff.

92
00:04:22,680 --> 00:04:24,010
Or we can make it
multiple choice.

93
00:04:24,010 --> 00:04:26,210
Well multiple choice,
we know it's easier.

94
00:04:26,210 --> 00:04:29,400
So all they have to do is cross
out which of these four

95
00:04:29,400 --> 00:04:31,770
things looks like this.

96
00:04:31,770 --> 00:04:33,360
And they picked the
circle here.

97
00:04:33,360 --> 00:04:35,610
Or instead of the x,
they picked the o.

98
00:04:35,610 --> 00:04:38,010
They're really not getting
almost anything about the

99
00:04:38,010 --> 00:04:39,140
quality of the shape.

100
00:04:39,140 --> 00:04:41,790
They're really not recognizing
what they're looking at.

101
00:04:41,790 --> 00:04:48,920
Or they'll see this.

102
00:04:48,920 --> 00:04:52,120
But they'll read it as, when
they get this, 7415.

103
00:04:52,120 --> 00:04:54,420
They won't take the first
interpretation, but will come

104
00:04:54,420 --> 00:04:55,670
up with some other.

105
00:04:58,250 --> 00:05:01,170
When these patients copy
something, they see, right in

106
00:05:01,170 --> 00:05:03,780
front of them, that elephant.

107
00:05:03,780 --> 00:05:04,920
You can label things.

108
00:05:04,920 --> 00:05:06,780
But it's a pretty bad copy.

109
00:05:06,780 --> 00:05:09,120
It's not just bad artistry.

110
00:05:09,120 --> 00:05:13,100
It's a bad sense of what the
coherent shape is right in

111
00:05:13,100 --> 00:05:13,600
front of them.

112
00:05:13,600 --> 00:05:15,510
And their vision is
basically fine.

113
00:05:15,510 --> 00:05:17,920
It's just the shape of it
doesn't make any sense.

114
00:05:21,710 --> 00:05:24,000
Not surprisingly then, maybe
these patients also struggle

115
00:05:24,000 --> 00:05:25,510
when you make things
a little harder.

116
00:05:25,510 --> 00:05:27,640
So what letter is this?

117
00:05:27,640 --> 00:05:28,400
You can get that.

118
00:05:28,400 --> 00:05:29,590
You'd have to look a
little bit harder.

119
00:05:29,590 --> 00:05:32,340
Or you might get that this is
a cow, or this is a car.

120
00:05:32,340 --> 00:05:34,610
These patients are terrible
at those things.

121
00:05:34,610 --> 00:05:35,630
As soon as you make
it a little bit

122
00:05:35,630 --> 00:05:39,120
hard, they're finished.

123
00:05:39,120 --> 00:05:41,400
They're also bad if they have
damage in the right

124
00:05:41,400 --> 00:05:43,460
hemisphere, posteriorly,
at what they

125
00:05:43,460 --> 00:05:44,770
call an unusual view.

126
00:05:44,770 --> 00:05:49,760
So often we see something like
a stapler from this view.

127
00:05:49,760 --> 00:05:51,770
We rarely quite see
it this way.

128
00:05:51,770 --> 00:05:53,720
But if I showed you this, you
could probably make it out.

129
00:05:53,720 --> 00:05:55,560
It might take you a
moment by itself.

130
00:05:55,560 --> 00:05:58,520
An unusual view of an object
that you see, psychologists

131
00:05:58,520 --> 00:06:01,280
call it a canonical view,
which is the usual one.

132
00:06:01,280 --> 00:06:03,130
And what happens to patients who
have damage in the right

133
00:06:03,130 --> 00:06:04,060
posterior cortex?

134
00:06:04,060 --> 00:06:08,680
They're terrible at making this
unusual perception of a

135
00:06:08,680 --> 00:06:10,090
common object.

136
00:06:10,090 --> 00:06:13,300
They're also terrible if they
have to see something that's a

137
00:06:13,300 --> 00:06:14,160
little bit challenging.

138
00:06:14,160 --> 00:06:17,280
Again, like here's some
hats but with shadows.

139
00:06:17,280 --> 00:06:19,090
Now normally, we don't
think about

140
00:06:19,090 --> 00:06:20,770
shadows as a big problem.

141
00:06:20,770 --> 00:06:23,590
It is for computers that
are meant to see.

142
00:06:23,590 --> 00:06:26,535
And it is also a big problem
for patients with damage to

143
00:06:26,535 --> 00:06:27,970
the right posterior cortex.

144
00:06:27,970 --> 00:06:30,590
It's as if they can't pull out
the object from the shadow.

145
00:06:30,590 --> 00:06:34,570
It's all one big glob to them.

146
00:06:34,570 --> 00:06:37,920
Compare that difficulty just
getting basically what's there

147
00:06:37,920 --> 00:06:40,180
to patients with left hemisphere
damage who can't

148
00:06:40,180 --> 00:06:41,750
recognize an object.

149
00:06:41,750 --> 00:06:45,160
So here's a patient who was
shown pictures of a key, a

150
00:06:45,160 --> 00:06:48,980
pig, a bird, a train, and
was asked to copy them.

151
00:06:48,980 --> 00:06:50,590
And the copies are actually
excellent.

152
00:06:50,590 --> 00:06:51,970
I mean, they're much better
than I would do.

153
00:06:51,970 --> 00:06:55,050
They're excellent
copies, right?

154
00:06:55,050 --> 00:06:57,100
But even after the patient has
made that excellent copy of

155
00:06:57,100 --> 00:06:58,530
the key, he's asked
what is that?

156
00:06:58,530 --> 00:07:00,970
And the patient says,
I don't know.

157
00:07:00,970 --> 00:07:04,130
Good copy of a pig, "Could be
a dog or any other animal."

158
00:07:04,130 --> 00:07:06,680
That's a dog, but that's it.

159
00:07:06,680 --> 00:07:08,120
The bird is kind
of interesting.

160
00:07:08,120 --> 00:07:10,050
He draws a pretty good bird,
his copy, right?

161
00:07:10,050 --> 00:07:13,540
But then he says, "It could be
a beach stump." And when you

162
00:07:13,540 --> 00:07:14,630
say that, you can see that.

163
00:07:14,630 --> 00:07:17,990
But you and I would never
come to that conclusion.

164
00:07:17,990 --> 00:07:19,860
And the train, he gets
a little bit of wagon

165
00:07:19,860 --> 00:07:20,840
or car of some kind.

166
00:07:20,840 --> 00:07:24,260
It's a larger being pulled
by a smaller.

167
00:07:24,260 --> 00:07:25,670
He vaguely knows
it's something.

168
00:07:25,670 --> 00:07:27,070
But it's not quite
right either.

169
00:07:27,070 --> 00:07:30,910
But he really gets what
it is, the shape.

170
00:07:30,910 --> 00:07:33,350
But it's as if it's stripped
of meaning.

171
00:07:33,350 --> 00:07:37,170
It just doesn't mean
anything to him.

172
00:07:37,170 --> 00:07:39,035
Now we're going to do
a video in a moment.

173
00:07:39,035 --> 00:07:41,320
All right, Tyler?

174
00:07:41,320 --> 00:07:42,580
I'm going to show you
some videos today.

175
00:07:42,580 --> 00:07:46,640
The problem is that these videos
are sort of grainy.

176
00:07:46,640 --> 00:07:49,650
And there's not really high
quality Alan Alda talking on

177
00:07:49,650 --> 00:07:52,550
Scientific American videos
of these patients.

178
00:07:52,550 --> 00:07:55,100
So the unfortunate thing is
in a visual disorders

179
00:07:55,100 --> 00:07:58,010
demonstration, it's almost as
if I'm forcing upon you that

180
00:07:58,010 --> 00:07:58,950
same visual disorder.

181
00:07:58,950 --> 00:07:59,140
OK?

182
00:07:59,140 --> 00:08:01,580
Because you're looking at it,
and you go, like, gee.

183
00:08:01,580 --> 00:08:02,840
No wonder he can't
recognize it.

184
00:08:02,840 --> 00:08:03,650
I can't recognize it.

185
00:08:03,650 --> 00:08:04,750
So let me tell you.

186
00:08:04,750 --> 00:08:07,110
In a couple moments he's
going to look at a

187
00:08:07,110 --> 00:08:08,930
picture of a clarinet.

188
00:08:08,930 --> 00:08:11,060
And if it were a really
good video, you

189
00:08:11,060 --> 00:08:11,800
would see it clearly.

190
00:08:11,800 --> 00:08:13,820
But there's just not many
of these around.

191
00:08:13,820 --> 00:08:15,870
And these patients are rare
that are this perfect.

192
00:08:15,870 --> 00:08:18,770
So I'm going to make
this dark.

193
00:08:18,770 --> 00:08:21,030
And we'll see-- thank
you, Tyler--

194
00:08:21,030 --> 00:08:24,060
a patient with visual agnosia.

195
00:08:24,060 --> 00:08:25,770
And you should be impressed
by two things.

196
00:08:25,770 --> 00:08:26,600
First, he's very verbal.

197
00:08:26,600 --> 00:08:27,280
He's speaking.

198
00:08:27,280 --> 00:08:28,350
He's smart.

199
00:08:28,350 --> 00:08:31,000
A lot of his brain is
working just fine.

200
00:08:31,000 --> 00:08:34,590
And second, you're going to
see something kind of

201
00:08:34,590 --> 00:08:36,940
interesting, which is
we talked about the

202
00:08:36,940 --> 00:08:38,580
where system last time.

203
00:08:38,580 --> 00:08:41,130
And sometimes people call
the where system,

204
00:08:41,130 --> 00:08:43,169
the system for action.

205
00:08:43,169 --> 00:08:45,580
Because pretty much like where
you grab something, where you

206
00:08:45,580 --> 00:08:48,060
move, that has a lot to do with
the action, the physical

207
00:08:48,060 --> 00:08:49,030
action in the world.

208
00:08:49,030 --> 00:08:52,050
So you're going to can see this
man unable to recognize

209
00:08:52,050 --> 00:08:55,120
by his what system, what
he's looking at.

210
00:08:55,120 --> 00:08:57,150
But a little bit of information
is going to sneak

211
00:08:57,150 --> 00:09:00,850
into his where system, and move
his hands in a way that

212
00:09:00,850 --> 00:09:01,600
you'll see appropriate.

213
00:09:01,600 --> 00:09:02,850
So thanks.

214
00:09:08,440 --> 00:09:11,240
So it's so funny if
you work in--

215
00:09:11,240 --> 00:09:14,250
this idea that we only
use 10% of our brain.

216
00:09:14,250 --> 00:09:19,010
I can assure you this man has
95%, 98% of his brain.

217
00:09:19,010 --> 00:09:19,615
He can speak.

218
00:09:19,615 --> 00:09:21,145
He can move.

219
00:09:21,145 --> 00:09:22,610
He can make jokes.

220
00:09:22,610 --> 00:09:24,230
He can curse at the end
there a little bit.

221
00:09:24,230 --> 00:09:26,810
He can do almost everything
that everybody can do.

222
00:09:26,810 --> 00:09:31,980
But he cannot identify what
an object is by vision.

223
00:09:31,980 --> 00:09:33,700
Because the injury has occurred
in the part of the

224
00:09:33,700 --> 00:09:36,120
brain that's essential for that
final step of figuring

225
00:09:36,120 --> 00:09:38,310
out what you're looking for.

226
00:09:38,310 --> 00:09:40,670
It's a very specific
part of the brain.

227
00:09:40,670 --> 00:09:44,490
And without it, the rest of
his brain can't figure out

228
00:09:44,490 --> 00:09:45,970
what it's looking at.

229
00:09:45,970 --> 00:09:47,260
So his sensation is fine.

230
00:09:47,260 --> 00:09:48,525
He sees where it is.

231
00:09:48,525 --> 00:09:51,020
And he gets the shape
even roughly.

232
00:09:51,020 --> 00:09:52,530
He thought it was a pen.

233
00:09:52,530 --> 00:09:55,290
There's something vaguely
about the shape.

234
00:09:55,290 --> 00:09:57,020
His naming is fine.

235
00:09:57,020 --> 00:10:01,060
But it's as if, again, this idea
of a perfect example of

236
00:10:01,060 --> 00:10:02,840
what we call associative
agnosia.

237
00:10:02,840 --> 00:10:04,990
He has all the shape information
it seems like you

238
00:10:04,990 --> 00:10:07,270
ought to have, all the language
information you have.

239
00:10:07,270 --> 00:10:10,170
But he can't make that move
from piecing together what

240
00:10:10,170 --> 00:10:12,350
something is to knowing
what it means.

241
00:10:12,350 --> 00:10:14,380
And if he doesn't know what
it means, he can't

242
00:10:14,380 --> 00:10:16,710
come up with the name.

243
00:10:16,710 --> 00:10:20,400
If he smells it or holds it,
he says, apple or pipe.

244
00:10:20,400 --> 00:10:21,640
He knows the name.

245
00:10:21,640 --> 00:10:24,450
But if he has to look at
something and then know what

246
00:10:24,450 --> 00:10:29,450
it is to name it, if I show you
a piece of paper and ask

247
00:10:29,450 --> 00:10:31,920
you to name it, you have
to say, what is it?

248
00:10:31,920 --> 00:10:33,640
Then I name it.

249
00:10:33,640 --> 00:10:38,140
He can't make that move in
vision to what it is.

250
00:10:38,140 --> 00:10:40,100
Because so much processing
has to occur in your

251
00:10:40,100 --> 00:10:41,030
brain to get there.

252
00:10:41,030 --> 00:10:42,590
Even though it seems
instantaneous

253
00:10:42,590 --> 00:10:43,840
and trivial to you.

254
00:10:46,380 --> 00:10:48,400
It's instantaneous and trivial
to you because your brain is

255
00:10:48,400 --> 00:10:52,050
so brilliant that it
accomplishes it.

256
00:10:55,570 --> 00:10:58,170
Now, there are things out
there in the world like

257
00:10:58,170 --> 00:11:00,130
animals that occur naturally.

258
00:11:00,130 --> 00:11:02,440
Because they're from an
evolutionary tree, they share

259
00:11:02,440 --> 00:11:03,960
a lot of shapes.

260
00:11:03,960 --> 00:11:06,660
And there's things out there
that are not living, either

261
00:11:06,660 --> 00:11:08,790
that occur like the moon and
the mountain-- they're out

262
00:11:08,790 --> 00:11:09,800
there sitting--

263
00:11:09,800 --> 00:11:12,700
or what we call manufactured
objects, things like locks,

264
00:11:12,700 --> 00:11:17,330
pencils, motorcycles, bulbs,
things that humans have made.

265
00:11:17,330 --> 00:11:20,620
And, of course, humans make
shapes as they need to

266
00:11:20,620 --> 00:11:21,730
construct things.

267
00:11:21,730 --> 00:11:24,920
They're not constrained by
evolutionary forces about what

268
00:11:24,920 --> 00:11:25,960
shape something is.

269
00:11:25,960 --> 00:11:29,610
We make a shape as we need it.

270
00:11:29,610 --> 00:11:32,220
But it was still an amazing
discovery in the 1980s that

271
00:11:32,220 --> 00:11:35,020
they found individual patients
who lost the ability to

272
00:11:35,020 --> 00:11:38,640
recognize things and various
specific categories.

273
00:11:38,640 --> 00:11:42,220
This seems so much like old
style, silly phrenology.

274
00:11:42,220 --> 00:11:45,450
That I can tell you, my
colleagues sort of make fun of

275
00:11:45,450 --> 00:11:48,990
me when I told them people in my
field were following this.

276
00:11:48,990 --> 00:11:53,190
So they found patients who
could define by pictures

277
00:11:53,190 --> 00:11:56,350
manufactured objects, things
like pencils, or typewriters,

278
00:11:56,350 --> 00:11:57,290
or computers.

279
00:11:57,290 --> 00:12:01,350
But they couldn't identify
foods or animals.

280
00:12:01,350 --> 00:12:03,700
And then they found patients
just the opposite who could

281
00:12:03,700 --> 00:12:07,870
identify foods and animals but
not manufactured objects.

282
00:12:07,870 --> 00:12:09,200
So that's really weird.

283
00:12:09,200 --> 00:12:12,870
That's as if you had organized
in your brain the manufactured

284
00:12:12,870 --> 00:12:14,380
section here like in
a supermarket.

285
00:12:14,380 --> 00:12:15,520
And here's the animal section.

286
00:12:15,520 --> 00:12:17,560
And here's the fruit section.

287
00:12:17,560 --> 00:12:19,660
Really, is that how our mind
is organized, like a

288
00:12:19,660 --> 00:12:21,690
supermarket?

289
00:12:21,690 --> 00:12:24,390
But then there were some
exceptions to these.

290
00:12:24,390 --> 00:12:25,740
And there are the exceptions
that helped us

291
00:12:25,740 --> 00:12:26,650
inform us on the rules.

292
00:12:26,650 --> 00:12:29,740
So here's some patients who are
really good at identifying

293
00:12:29,740 --> 00:12:32,740
manufactured things-- again,
pencils, things like that--

294
00:12:32,740 --> 00:12:35,126
bad at foods and animals.

295
00:12:35,126 --> 00:12:36,230
But now think about this.

296
00:12:36,230 --> 00:12:39,030
They were good on body parts.

297
00:12:39,030 --> 00:12:41,290
How were they good on body parts
when they were bad at

298
00:12:41,290 --> 00:12:41,830
foods and animals?

299
00:12:41,830 --> 00:12:44,950
Body parts are natural, right?

300
00:12:44,950 --> 00:12:47,750
And they were bad a musical
instruments even though they

301
00:12:47,750 --> 00:12:50,310
were good on other manufactured
things like

302
00:12:50,310 --> 00:12:52,850
pencils and computers.

303
00:12:52,850 --> 00:12:54,910
So these were two
wrong things.

304
00:12:54,910 --> 00:12:57,270
And I'm going to try to convince
in a moment that the

305
00:12:57,270 --> 00:13:00,270
way we now understand this is
they're good at body parts

306
00:13:00,270 --> 00:13:02,750
because we know how
to use our body.

307
00:13:02,750 --> 00:13:06,160
And they're bad at musical
instruments because musical

308
00:13:06,160 --> 00:13:10,500
instruments look an awful lot
like one another just like

309
00:13:10,500 --> 00:13:13,280
four-legged animals look
a lot like one another.

310
00:13:13,280 --> 00:13:15,810
That's a funny category for us
if we start to divide up

311
00:13:15,810 --> 00:13:17,330
instruments.

312
00:13:17,330 --> 00:13:23,260
So the problem is not so much
one of natural versus

313
00:13:23,260 --> 00:13:24,610
unnatural things.

314
00:13:24,610 --> 00:13:27,960
It's things that we know through
vision versus things

315
00:13:27,960 --> 00:13:30,690
that we know through physical
action and use.

316
00:13:30,690 --> 00:13:31,640
And here's another example.

317
00:13:31,640 --> 00:13:37,970
The patients who are bad at
objects that we use every day,

318
00:13:37,970 --> 00:13:40,420
like pencils and stuff like
that, were pretty good on

319
00:13:40,420 --> 00:13:43,320
things like billboards
and mountains.

320
00:13:43,320 --> 00:13:47,140
Now only in science fiction
movies do you have creatures

321
00:13:47,140 --> 00:13:48,770
big enough that they're picking
up billboards and

322
00:13:48,770 --> 00:13:50,400
throwing it at helicopters
from the military.

323
00:13:50,400 --> 00:13:51,940
You see that all the time.

324
00:13:51,940 --> 00:13:52,530
You have the big creature.

325
00:13:52,530 --> 00:13:53,800
And they pick it up,
and they hurl it.

326
00:13:53,800 --> 00:13:54,410
OK.

327
00:13:54,410 --> 00:14:00,110
So we don't mostly pick up huge
things like mountains and

328
00:14:00,110 --> 00:14:02,380
large outdoor objects
like flag poles.

329
00:14:02,380 --> 00:14:06,240
They're not things that
we pick up and use.

330
00:14:06,240 --> 00:14:09,320
So it seems like the principle
of what's good or bad is more

331
00:14:09,320 --> 00:14:13,080
about how we experience things
through sight or through

332
00:14:13,080 --> 00:14:14,940
physical action.

333
00:14:14,940 --> 00:14:17,280
So how do we know things in
the world through visual

334
00:14:17,280 --> 00:14:19,410
experience?

335
00:14:19,410 --> 00:14:21,160
Unless we're alligator
wrestlers, we don't get that

336
00:14:21,160 --> 00:14:23,540
near that many animals.

337
00:14:23,540 --> 00:14:26,660
And most of us have other things
that we use all the

338
00:14:26,660 --> 00:14:29,860
time, like bulbs, and pencils,
and things like that.

339
00:14:29,860 --> 00:14:31,350
And evidence for
that was this.

340
00:14:31,350 --> 00:14:34,190
They put humans inside a scanner
and looked at their

341
00:14:34,190 --> 00:14:37,440
brains as they named either
line drawings or words.

342
00:14:37,440 --> 00:14:38,960
So let's think of words.

343
00:14:38,960 --> 00:14:42,730
So it's just words, and you're
naming things like elephant

344
00:14:42,730 --> 00:14:43,540
and tiger--

345
00:14:43,540 --> 00:14:44,600
that's the animal--

346
00:14:44,600 --> 00:14:47,670
or tools that you use, like
pliers or hammer.

347
00:14:47,670 --> 00:14:49,330
Visually, they're words.

348
00:14:49,330 --> 00:14:52,130
But what happens in the brain
as you're naming animals?

349
00:14:52,130 --> 00:14:54,090
It seems like you turn on
the visual parts of

350
00:14:54,090 --> 00:14:54,800
your brain a lot.

351
00:14:54,800 --> 00:14:56,170
There's no animal in there.

352
00:14:56,170 --> 00:14:56,980
There's not even a picture
of an animal.

353
00:14:56,980 --> 00:15:00,630
It's just the word elephant
or the word cat.

354
00:15:00,630 --> 00:15:02,860
What happens when you're
naming tools?

355
00:15:02,860 --> 00:15:05,820
In the left hemisphere, we see
turned on an area here that's

356
00:15:05,820 --> 00:15:11,240
involved in visual motion, and
here, in physical action of

357
00:15:11,240 --> 00:15:13,740
moving things in the world.

358
00:15:13,740 --> 00:15:15,900
All you're doing is
naming words.

359
00:15:15,900 --> 00:15:20,050
But it's as if the knowledge for
naming them is located in

360
00:15:20,050 --> 00:15:23,560
the part of your brain that
uses that information.

361
00:15:23,560 --> 00:15:27,410
If there's things I recognize
by sight, the abstract

362
00:15:27,410 --> 00:15:30,350
knowledge about those things
is near my visual system.

363
00:15:30,350 --> 00:15:34,360
If there's things I recognize
by action, using them, the

364
00:15:34,360 --> 00:15:37,100
abstract knowledge of that is
located in your brain near

365
00:15:37,100 --> 00:15:39,710
parts of your brain that move
things and pick up things.

366
00:15:39,710 --> 00:15:41,760
So all of a sudden, it makes
sense that you can get

367
00:15:41,760 --> 00:15:43,360
separation in these forms
of knowledge.

368
00:15:43,360 --> 00:15:46,310
Because they reside in your
brain near the parts of your

369
00:15:46,310 --> 00:15:49,400
brain that use them and
communicate with them.

370
00:15:49,400 --> 00:15:51,160
So again, this idea that if
it's a [INAUDIBLE] like a

371
00:15:51,160 --> 00:15:54,775
scissors, your knowledge about
scissors will reside in the

372
00:15:54,775 --> 00:15:57,300
parts of the brain that
act upon scissors.

373
00:15:57,300 --> 00:16:01,540
If it's something like a bird
that you mostly see, knowledge

374
00:16:01,540 --> 00:16:04,540
about that bird, not just how it
looks, but everything, will

375
00:16:04,540 --> 00:16:08,120
reside in parts of the brain
that are focused on vision.

376
00:16:08,120 --> 00:16:09,940
So it's as if your knowledge
is moved to the part of the

377
00:16:09,940 --> 00:16:14,040
brain that interacts with
the form of knowledge.

378
00:16:14,040 --> 00:16:15,360
So now let's talk about faces.

379
00:16:15,360 --> 00:16:19,170
Faces are a huge thing for us.

380
00:16:19,170 --> 00:16:22,370
There's at least two huge
things we do with faces.

381
00:16:22,370 --> 00:16:26,140
We recognize who people are
and what their moods are,

382
00:16:26,140 --> 00:16:29,760
angry, happy, nice, and so on.

383
00:16:29,760 --> 00:16:31,500
So here's one person.

384
00:16:31,500 --> 00:16:32,660
Here's another person.

385
00:16:32,660 --> 00:16:34,870
When I first used this picture,
he wasn't a governor.

386
00:16:34,870 --> 00:16:35,570
He was a movie star.

387
00:16:35,570 --> 00:16:37,690
And now he's an ex-governor.

388
00:16:37,690 --> 00:16:39,250
And you have the problem
that you have to

389
00:16:39,250 --> 00:16:40,630
recognize a lot of people.

390
00:16:40,630 --> 00:16:42,440
Because you have all your
classmates, and friends, and

391
00:16:42,440 --> 00:16:44,930
family members, and a lot
of people out there.

392
00:16:44,930 --> 00:16:46,670
And not only that, you have
to recognize the various

393
00:16:46,670 --> 00:16:48,940
expressions they have, which are
terribly important for how

394
00:16:48,940 --> 00:16:53,900
we communicate with people
moment to moment all the time.

395
00:16:53,900 --> 00:16:56,120
So we could say with faces,
our big job is who are we

396
00:16:56,120 --> 00:16:57,260
looking at.

397
00:16:57,260 --> 00:17:01,420
A friend, or foe, somebody
loved, somebody we don't know.

398
00:17:01,420 --> 00:17:02,600
Who is that person?

399
00:17:02,600 --> 00:17:04,609
And then what's the expression
that we can

400
00:17:04,609 --> 00:17:05,470
read from their face?

401
00:17:05,470 --> 00:17:08,140
That's a huge part of our lives
as social animals, as

402
00:17:08,140 --> 00:17:11,300
people who interact
with other people.

403
00:17:11,300 --> 00:17:16,460
So you saw a film last time of
a patient with prosopagnosia.

404
00:17:16,460 --> 00:17:18,940
Your book, this time, talks
about a Dr. P. He's the man

405
00:17:18,940 --> 00:17:20,829
who mistook his wife
for a hat.

406
00:17:20,829 --> 00:17:23,810
So he has bigger problems in
vision than simply faces.

407
00:17:23,810 --> 00:17:25,500
But faces are very impressive.

408
00:17:25,500 --> 00:17:28,120
He fails to recognize his
students by their face.

409
00:17:28,120 --> 00:17:30,020
But he can recognize them
by their voice,

410
00:17:30,020 --> 00:17:31,450
modality-specific.

411
00:17:31,450 --> 00:17:34,080
His general vision is OK.

412
00:17:34,080 --> 00:17:37,840
When Oliver Sacks visited him
in his house, he tries to

413
00:17:37,840 --> 00:17:40,950
reach out his hand and took hold
of his wife's head, tried

414
00:17:40,950 --> 00:17:42,585
to lift it off to put it on.

415
00:17:42,585 --> 00:17:46,880
He had apparently mistaken
his wife for a hat.

416
00:17:46,880 --> 00:17:49,440
His wife looked as if she
was used to such things.

417
00:17:49,440 --> 00:17:51,260
You can see the head and the
hat, they're kind of close.

418
00:17:51,260 --> 00:17:53,560
But not if you have
typical vision.

419
00:17:53,560 --> 00:17:56,490
He might pat the head of water
hydrants and parking meters

420
00:17:56,490 --> 00:17:59,740
mistaking them for the
heads of children.

421
00:17:59,740 --> 00:18:01,190
There's a little bit of logic.

422
00:18:01,190 --> 00:18:03,315
But it's pretty wrong
what he sees.

423
00:18:03,315 --> 00:18:04,765
And it's pretty meaningless
to him.

424
00:18:07,380 --> 00:18:09,850
So let me show you a
film of an example.

425
00:18:09,850 --> 00:18:11,100
You saw one last time.

426
00:18:13,810 --> 00:18:14,750
So let me say a word
about this.

427
00:18:14,750 --> 00:18:16,920
Imagine that if you couldn't
recognize people

428
00:18:16,920 --> 00:18:17,770
as you walked up.

429
00:18:17,770 --> 00:18:20,960
So much of how we recognize
people-- we don't think about

430
00:18:20,960 --> 00:18:22,930
this, because we instantly,
easily recognize

431
00:18:22,930 --> 00:18:24,050
those we know well--

432
00:18:24,050 --> 00:18:26,040
is by their faces.

433
00:18:26,040 --> 00:18:28,390
So I want to say a
couple of things.

434
00:18:28,390 --> 00:18:29,990
Usually there's posterior
cortical lesions

435
00:18:29,990 --> 00:18:31,710
like in this woman.

436
00:18:31,710 --> 00:18:34,800
I knew one man who was a
Unitarian minister, who I

437
00:18:34,800 --> 00:18:36,760
worked with, actually, a couple
blocks away here,

438
00:18:36,760 --> 00:18:39,120
He went up to Dartmouth for
the winter carnival.

439
00:18:39,120 --> 00:18:40,960
He came back in a small
Volkswagen.

440
00:18:40,960 --> 00:18:41,870
It turned over turned
over on the ice.

441
00:18:41,870 --> 00:18:45,505
He had a huge brain injury
and became prosopagnosic.

442
00:18:45,505 --> 00:18:47,900
He could only recognize his wife
by asking her to wear a

443
00:18:47,900 --> 00:18:49,700
yellow bow in her hair.

444
00:18:49,700 --> 00:18:50,660
Because otherwise he
couldn't pick her

445
00:18:50,660 --> 00:18:52,370
out from other people.

446
00:18:52,370 --> 00:18:55,740
When he went down the street
and passed his brother, he

447
00:18:55,740 --> 00:18:57,410
completely failed to
recognize him.

448
00:18:57,410 --> 00:18:58,510
Then the brother spoke.

449
00:18:58,510 --> 00:18:59,420
Then he recognized him.

450
00:18:59,420 --> 00:19:01,740
And he said-- and this was his
subjective experience--

451
00:19:01,740 --> 00:19:03,400
because you wonder what
do these people

452
00:19:03,400 --> 00:19:05,670
feel like they see.

453
00:19:05,670 --> 00:19:08,810
He said that what it felt like
for his brother was, do you

454
00:19:08,810 --> 00:19:11,340
when somebody has some bad
handwriting, and you can't

455
00:19:11,340 --> 00:19:12,480
make it out at all?

456
00:19:12,480 --> 00:19:13,830
And then somebody tells
you what it is.

457
00:19:13,830 --> 00:19:15,280
And you go, oh yes.

458
00:19:15,280 --> 00:19:16,260
That's that word.

459
00:19:16,260 --> 00:19:17,730
Any now you see it.

460
00:19:17,730 --> 00:19:20,120
He said, that's what it felt
like as he heard his brother's

461
00:19:20,120 --> 00:19:22,990
voice and realized it
was his brother.

462
00:19:22,990 --> 00:19:24,320
It's hard to know how to
relate these things.

463
00:19:24,320 --> 00:19:27,180
But the ability to recognize a
person is completely wiped out

464
00:19:27,180 --> 00:19:28,610
of these individuals.

465
00:19:28,610 --> 00:19:32,270
In the last six or seven years,
something else has been

466
00:19:32,270 --> 00:19:32,570
discovered.

467
00:19:32,570 --> 00:19:34,570
Now let me tell you
a word about this.

468
00:19:34,570 --> 00:19:37,560
Instead of people with acute
lesions that they got because

469
00:19:37,560 --> 00:19:38,810
of a brain injury.

470
00:19:38,810 --> 00:19:42,440
It turns out there's a couple
percent of our population

471
00:19:42,440 --> 00:19:45,590
that's really bad at recognizing
faces, not the

472
00:19:45,590 --> 00:19:47,630
occasional social embarrassment
that many of us

473
00:19:47,630 --> 00:19:50,880
have when we see somebody, we
know we should know them, but

474
00:19:50,880 --> 00:19:53,560
they're really, really,
really bad.

475
00:19:53,560 --> 00:19:56,250
There's websites you can go to
that let you test yourself.

476
00:19:56,250 --> 00:19:58,570
Because researchers want you
to work with them if you're

477
00:19:58,570 --> 00:19:59,800
one of them.

478
00:19:59,800 --> 00:20:02,500
It's about 1% of
the population.

479
00:20:02,500 --> 00:20:05,200
They're really, really,
really bad.

480
00:20:05,200 --> 00:20:06,480
There was one woman described
as, she says,

481
00:20:06,480 --> 00:20:07,690
they can't watch movies.

482
00:20:07,690 --> 00:20:09,070
Because all the characters
look pretty

483
00:20:09,070 --> 00:20:12,320
much the same to them.

484
00:20:12,320 --> 00:20:13,570
You can imagine that.

485
00:20:16,050 --> 00:20:19,600
One parent who has this problem,
really terrible at

486
00:20:19,600 --> 00:20:22,480
faces by formal testing, says
that when her kid comes out of

487
00:20:22,480 --> 00:20:26,800
school, her own kid, if he is
not wearing the clothes that

488
00:20:26,800 --> 00:20:29,270
she knows he's wearing, she
can't pick him out from the

489
00:20:29,270 --> 00:20:31,300
other children very easily.

490
00:20:31,300 --> 00:20:33,390
And these are not people with
big brain injuries.

491
00:20:33,390 --> 00:20:35,820
These are people who have some
unusual difficulty in

492
00:20:35,820 --> 00:20:36,870
recognizing faces.

493
00:20:36,870 --> 00:20:40,440
So in a classroom this
size, there easily

494
00:20:40,440 --> 00:20:43,200
could be one of you.

495
00:20:43,200 --> 00:20:44,640
We don't understand
what that is.

496
00:20:44,640 --> 00:20:46,740
It's not a brain injury.

497
00:20:46,740 --> 00:20:49,490
Just like some of us can't carry
a tune, some individuals

498
00:20:49,490 --> 00:20:54,390
are really terrible at
recognizing faces, really bad.

499
00:20:54,390 --> 00:20:56,560
So there's a lot of research
into that to understand what

500
00:20:56,560 --> 00:20:57,740
that reflects.

501
00:20:57,740 --> 00:20:59,870
So let me tell you what we do
understand about the neural

502
00:20:59,870 --> 00:21:02,680
circuitry in the typical brain
for recognizing a face of a

503
00:21:02,680 --> 00:21:04,820
loved one, a person that
you care about.

504
00:21:04,820 --> 00:21:06,960
So this is the fusiform gyrus.

505
00:21:06,960 --> 00:21:09,060
It's a structure in the back of
the brain that extends from

506
00:21:09,060 --> 00:21:11,170
the occipital into the
temporal lobes.

507
00:21:11,170 --> 00:21:13,620
And that's an anatomical
description of this region.

508
00:21:13,620 --> 00:21:18,540
Within that region, there's a
functional area that seems

509
00:21:18,540 --> 00:21:20,660
very important for
face perception.

510
00:21:20,660 --> 00:21:22,930
And Nancy Kanwisher, who's in
our department, pretty much

511
00:21:22,930 --> 00:21:25,900
discovered this specific
functional area in the typical

512
00:21:25,900 --> 00:21:27,250
human brain.

513
00:21:27,250 --> 00:21:30,340
And so she did experiments where
she would present things

514
00:21:30,340 --> 00:21:32,110
like faces or objects.

515
00:21:35,700 --> 00:21:37,870
In this picture, left is right,
and right is left.

516
00:21:37,870 --> 00:21:40,270
There's a spot in the fusiform
cortex that has

517
00:21:40,270 --> 00:21:41,580
this following property.

518
00:21:41,580 --> 00:21:42,280
Here's faces.

519
00:21:42,280 --> 00:21:43,070
That turns on.

520
00:21:43,070 --> 00:21:45,065
Here's objects which have
similar properties.

521
00:21:45,065 --> 00:21:47,530
It has a lot of visual
information like a face.

522
00:21:47,530 --> 00:21:48,740
But it's not a face.

523
00:21:48,740 --> 00:21:51,420
It's part of the brain region
not so interested, faces,

524
00:21:51,420 --> 00:21:53,310
objects, faces, objects.

525
00:21:53,310 --> 00:21:54,970
So whenever faces come on--

526
00:21:54,970 --> 00:21:56,860
this is, by our understanding--
the first part

527
00:21:56,860 --> 00:21:59,630
of your brain that says,
I'm looking at a face.

528
00:21:59,630 --> 00:22:02,130
Up until then, it's just treated
like other stuff.

529
00:22:02,130 --> 00:22:04,710
This is the first part of your
brain that recognizes, I'm

530
00:22:04,710 --> 00:22:05,870
looking at a face.

531
00:22:05,870 --> 00:22:08,890
I'm starting to identify who
I'm looking at and what

532
00:22:08,890 --> 00:22:11,950
feelings they might have
by their expression.

533
00:22:11,950 --> 00:22:14,970
And Nancy and her colleagues did
all kinds of very clever

534
00:22:14,970 --> 00:22:18,450
experiments to show the
properties of this.

535
00:22:18,450 --> 00:22:19,450
Let me pick one.

536
00:22:19,450 --> 00:22:21,970
You could say, well, it's
not about faces.

537
00:22:21,970 --> 00:22:23,420
But faces are about humans.

538
00:22:23,420 --> 00:22:25,550
Maybe it's a part of the brain
that tells you when there's a

539
00:22:25,550 --> 00:22:26,970
human around you.

540
00:22:26,970 --> 00:22:29,870
So they compared faces
versus hands.

541
00:22:29,870 --> 00:22:30,670
Here's hands.

542
00:22:30,670 --> 00:22:32,060
This part of the brain is
not that interested.

543
00:22:32,060 --> 00:22:33,900
Faces, hands, faces, hands.

544
00:22:33,900 --> 00:22:35,770
You see every time it moves to
faces, this part of the brain

545
00:22:35,770 --> 00:22:37,280
gets very strongly engaged.

546
00:22:37,280 --> 00:22:42,330
It's very much about faces
by many research studies.

547
00:22:42,330 --> 00:22:44,030
So let me tell you a little bit
about what we understand

548
00:22:44,030 --> 00:22:45,750
about faces in terms
of development

549
00:22:45,750 --> 00:22:46,430
and things like that.

550
00:22:46,430 --> 00:22:48,920
Because faces are such a big
part of our lives from the

551
00:22:48,920 --> 00:22:50,420
first few moments we're born.

552
00:22:50,420 --> 00:22:55,870
In fact, a colleague of mine,
Pawan Sinha, did a kind of a

553
00:22:55,870 --> 00:22:58,760
biographical research study that
was on the front page of

554
00:22:58,760 --> 00:23:02,030
The New York Times and earned
MIT faculty the continuing

555
00:23:02,030 --> 00:23:03,560
names of being Frankenstein
faculty.

556
00:23:03,560 --> 00:23:05,240
There was a number of faculty.

557
00:23:05,240 --> 00:23:08,400
And what he did, kind of very
cleverly, was he put on his

558
00:23:08,400 --> 00:23:12,360
newborn infant's head
a little camera.

559
00:23:12,360 --> 00:23:14,150
And he said, if I leave
it on there and

560
00:23:14,150 --> 00:23:15,240
record it all the time--

561
00:23:15,240 --> 00:23:16,950
and apparently there was some
debate with his wife whether

562
00:23:16,950 --> 00:23:17,770
he was allowed to do this.

563
00:23:17,770 --> 00:23:18,410
But they checked it out.

564
00:23:18,410 --> 00:23:21,640
It seemed OK by the
pediatrician.

565
00:23:21,640 --> 00:23:24,020
He said, if I have the camera
on the infant's head, I can

566
00:23:24,020 --> 00:23:27,050
see what the infant's experience
is like moment to

567
00:23:27,050 --> 00:23:30,180
moment in the first months
of its life.

568
00:23:30,180 --> 00:23:34,350
And I can also play with the
video so that it has, roughly

569
00:23:34,350 --> 00:23:39,540
speaking, the properties of the
infant's visual system,

570
00:23:39,540 --> 00:23:42,955
which is it can't see very
far after its birth.

571
00:23:42,955 --> 00:23:45,070
They can only see very close
by, and not that well.

572
00:23:45,070 --> 00:23:46,333
Over the next several months--
we'll talk about

573
00:23:46,333 --> 00:23:46,910
this later in class--

574
00:23:46,910 --> 00:23:49,860
he gets better and better
at seeing at a distance.

575
00:23:49,860 --> 00:23:53,230
So here's the infant sitting
in his or her crib.

576
00:23:53,230 --> 00:23:56,000
And this video makes incredibly
clear, beautifully

577
00:23:56,000 --> 00:23:58,900
and fascinatingly clear,
something that I never thought

578
00:23:58,900 --> 00:24:02,140
of until Pawan did this study,
this examination.

579
00:24:02,140 --> 00:24:04,800
So the infant sitting there,
it can't move very much.

580
00:24:04,800 --> 00:24:07,480
It can't go very many places
as a newborn infant.

581
00:24:07,480 --> 00:24:08,410
So what does it see?

582
00:24:08,410 --> 00:24:12,630
What gets close enough to it to
be visible to the infant?

583
00:24:12,630 --> 00:24:14,430
People's faces.

584
00:24:14,430 --> 00:24:16,825
Because all the other cute
things in the nursery, all the

585
00:24:16,825 --> 00:24:19,070
other things that are there,
they are too far or too blurry

586
00:24:19,070 --> 00:24:21,410
for the infant to process
almost at all.

587
00:24:21,410 --> 00:24:24,290
Almost the only thing that
happens is the parent, the

588
00:24:24,290 --> 00:24:27,430
grandparent, the siblings, the
friends that come over and go,

589
00:24:27,430 --> 00:24:28,610
what a cute baby.

590
00:24:28,610 --> 00:24:29,870
Baby, baby, baby, baby.

591
00:24:29,870 --> 00:24:32,190
The face comes in.

592
00:24:32,190 --> 00:24:34,430
The person's done, gone.

593
00:24:34,430 --> 00:24:36,350
Now you're just sitting
in a world of blur.

594
00:24:36,350 --> 00:24:38,790
Here comes another face.

595
00:24:38,790 --> 00:24:41,090
So besides certain issues
related to perhaps the

596
00:24:41,090 --> 00:24:43,430
feeding, the only thing
the infant sees are

597
00:24:43,430 --> 00:24:45,660
faces all the time.

598
00:24:45,660 --> 00:24:49,460
So it's a very important
social cue of who it's

599
00:24:49,460 --> 00:24:50,600
interacting with and
what's out there.

600
00:24:50,600 --> 00:24:52,680
And it's almost the only
one available to it.

601
00:24:52,680 --> 00:24:56,330
So psychologists have known for
some time that if you show

602
00:24:56,330 --> 00:24:59,850
a very young infant a face
versus something like an

603
00:24:59,850 --> 00:25:02,430
object, like a car or anything
like that, they're much more

604
00:25:02,430 --> 00:25:04,980
interested in faces if you put
them next to each other where

605
00:25:04,980 --> 00:25:06,440
the infant can see them.

606
00:25:06,440 --> 00:25:08,330
And they've been saying,
well, is the brain

607
00:25:08,330 --> 00:25:10,130
born to process faces?

608
00:25:10,130 --> 00:25:10,770
And it might be.

609
00:25:10,770 --> 00:25:12,100
And I'll show you evidence
to suggest in some

610
00:25:12,100 --> 00:25:13,510
general sense it is.

611
00:25:13,510 --> 00:25:16,290
What is that infant's
brain drawn to?

612
00:25:16,290 --> 00:25:19,100
What does it find fascinating
to learn about in the world?

613
00:25:19,100 --> 00:25:22,370
And so they notice that one
thing about a face is in terms

614
00:25:22,370 --> 00:25:27,720
of features, eyes, nose,
mouth, top-heavy.

615
00:25:27,720 --> 00:25:29,800
Because we have two eyes, and
everything else is one and

616
00:25:29,800 --> 00:25:31,470
one, the nose and the mouth.

617
00:25:31,470 --> 00:25:33,750
So here you can move the
features like this.

618
00:25:33,750 --> 00:25:36,020
And it turns out not only are
infants more interested in

619
00:25:36,020 --> 00:25:38,300
this kind of face than this--
and they know that because

620
00:25:38,300 --> 00:25:40,390
they'll put the two up, and
they'll measure how long the

621
00:25:40,390 --> 00:25:43,760
infant looks at one versus the
other, which is interesting.

622
00:25:43,760 --> 00:25:44,730
And the infant picks this one.

623
00:25:44,730 --> 00:25:45,590
But get a load of this?

624
00:25:45,590 --> 00:25:48,540
The infant also picks this one
versus this, this one versus

625
00:25:48,540 --> 00:25:50,200
this, and this one
versus this.

626
00:25:50,200 --> 00:25:52,710
Now these look very little
like a face.

627
00:25:52,710 --> 00:25:55,490
But they're top-heavy.

628
00:25:55,490 --> 00:25:59,500
It's as if maybe the infant's
brain to start with, maybe,

629
00:25:59,500 --> 00:26:02,380
looks for some interesting,
top-heavy things and that

630
00:26:02,380 --> 00:26:04,640
pulls it into the
world of faces.

631
00:26:04,640 --> 00:26:06,500
Here's some more experiments
like that.

632
00:26:06,500 --> 00:26:12,040
So here's an upside down,
right-side up.

633
00:26:12,040 --> 00:26:13,960
Baby looks more at this
kind of face.

634
00:26:13,960 --> 00:26:14,240
OK.

635
00:26:14,240 --> 00:26:15,790
You could say, well, that's
a normal face.

636
00:26:15,790 --> 00:26:17,840
But it's also top-heavy.

637
00:26:17,840 --> 00:26:21,900
Then they'd make these kind
of scary-looking faces.

638
00:26:21,900 --> 00:26:24,860
And now the baby, even though
these are nothing like a face

639
00:26:24,860 --> 00:26:28,540
the baby ever sees, it
still likes this one

640
00:26:28,540 --> 00:26:31,260
because it's top-heavy.

641
00:26:31,260 --> 00:26:32,740
Now how about this one?

642
00:26:32,740 --> 00:26:33,930
This is like a face they see.

643
00:26:33,930 --> 00:26:36,110
This is a really weird face.

644
00:26:36,110 --> 00:26:36,490
Right?

645
00:26:36,490 --> 00:26:38,020
They'll never see a face like
this except in this

646
00:26:38,020 --> 00:26:39,160
experiment.

647
00:26:39,160 --> 00:26:40,150
Top-heavy.

648
00:26:40,150 --> 00:26:41,180
Top-heavy.

649
00:26:41,180 --> 00:26:43,700
And now the infant looks at
both of them with equal

650
00:26:43,700 --> 00:26:45,970
interest and delight.

651
00:26:45,970 --> 00:26:49,860
So that's a very clever
experiment to suggest that the

652
00:26:49,860 --> 00:26:53,500
infant's mind is looking
for top-heavy things.

653
00:26:53,500 --> 00:26:55,040
And you could say, is
it really top-heavy?

654
00:26:55,040 --> 00:26:58,260
Or Is that what the infant's
mind and brain can look for?

655
00:26:58,260 --> 00:26:59,480
And it goes with faces.

656
00:26:59,480 --> 00:27:01,830
And evolution picked that.

657
00:27:01,830 --> 00:27:03,850
It's hard to know how to think
about these things.

658
00:27:03,850 --> 00:27:06,500
But it looks like the
top-heaviness draws an infant

659
00:27:06,500 --> 00:27:08,960
to be fascinated by faces.

660
00:27:08,960 --> 00:27:12,880
And luckily, you don't get
many faces like this to

661
00:27:12,880 --> 00:27:16,920
confuse you in the real world.

662
00:27:16,920 --> 00:27:18,120
Another thing that people
have studied a lot

663
00:27:18,120 --> 00:27:21,590
about faces is this.

664
00:27:21,590 --> 00:27:23,040
They use the word configural.

665
00:27:23,040 --> 00:27:25,230
The relations among the parts
are more important than the

666
00:27:25,230 --> 00:27:26,230
parts themselves.

667
00:27:26,230 --> 00:27:27,540
So I need a volunteer
in their seat.

668
00:27:27,540 --> 00:27:29,850
And this is going to be really
easy if you know the alphabet.

669
00:27:32,380 --> 00:27:35,270
You have to put up your
hand and help me out.

670
00:27:35,270 --> 00:27:35,490
OK.

671
00:27:35,490 --> 00:27:36,240
Thank you very much.

672
00:27:36,240 --> 00:27:38,410
I'm going to describe to you
a letter by its features.

673
00:27:38,410 --> 00:27:39,190
OK?

674
00:27:39,190 --> 00:27:41,470
It has a big line at the top.

675
00:27:41,470 --> 00:27:42,510
And in the middle, there's
a line that

676
00:27:42,510 --> 00:27:44,150
goes down to the bottom.

677
00:27:44,150 --> 00:27:45,140
What's the letter?

678
00:27:45,140 --> 00:27:45,610
AUDIENCE: T.

679
00:27:45,610 --> 00:27:45,980
PROFESSOR: Awesome.

680
00:27:45,980 --> 00:27:46,430
OK.

681
00:27:46,430 --> 00:27:47,490
Featural description.

682
00:27:47,490 --> 00:27:50,600
Now I'm going to describe to you
a super famous movie star.

683
00:27:50,600 --> 00:27:53,790
And this person has two eyes to
the left and to the right,

684
00:27:53,790 --> 00:27:56,150
a nose right below that, and
a mouth right below that.

685
00:27:56,150 --> 00:27:59,450
And who is it?

686
00:27:59,450 --> 00:28:02,290
Angelina Jolie.

687
00:28:02,290 --> 00:28:03,390
So what's my point here?

688
00:28:03,390 --> 00:28:06,210
It's the stuff that helps us
in many things, like the

689
00:28:06,210 --> 00:28:06,770
letters-- thank you.

690
00:28:06,770 --> 00:28:09,240
That was excellent.

691
00:28:09,240 --> 00:28:11,010
We can describe those
by the parts.

692
00:28:11,010 --> 00:28:12,560
And the parts pretty much
define the whole.

693
00:28:12,560 --> 00:28:13,640
That's not true for faces.

694
00:28:13,640 --> 00:28:15,930
It really is the precise
relations between eyes, and

695
00:28:15,930 --> 00:28:17,210
nose, and mouth that tell us.

696
00:28:17,210 --> 00:28:20,120
And there's an experiment that
shows this in a very empirical

697
00:28:20,120 --> 00:28:21,030
direct way.

698
00:28:21,030 --> 00:28:22,870
So they would show people
a face like this with a

699
00:28:22,870 --> 00:28:25,420
particular nose or a house
like this with

700
00:28:25,420 --> 00:28:27,440
a particular door.

701
00:28:27,440 --> 00:28:30,000
And then they would test your
memory for that nose or that

702
00:28:30,000 --> 00:28:35,910
door by itself as features or
with the whole face or a whole

703
00:28:35,910 --> 00:28:37,840
house present.

704
00:28:37,840 --> 00:28:40,920
And here's what happens for
people's performance.

705
00:28:40,920 --> 00:28:44,570
The nose by itself, the door by
itself, the nose when you

706
00:28:44,570 --> 00:28:46,370
see the whole picture, which
could be his nose.

707
00:28:46,370 --> 00:28:47,650
It could be another nose.

708
00:28:47,650 --> 00:28:48,290
Here's the door.

709
00:28:48,290 --> 00:28:50,310
It could be that door
or a different door.

710
00:28:50,310 --> 00:28:52,520
And here's what happens
in performance.

711
00:28:52,520 --> 00:28:56,560
If it's houses, people are good
for doors, about the same

712
00:28:56,560 --> 00:28:59,880
whether they've identified the
door by itself or the door in

713
00:28:59,880 --> 00:29:01,980
the context of the house.

714
00:29:01,980 --> 00:29:04,540
If it's noses, they were
a lot better off

715
00:29:04,540 --> 00:29:06,100
seeing the entire face.

716
00:29:06,100 --> 00:29:10,080
It's as if how we analyzed even
the parts of the face are

717
00:29:10,080 --> 00:29:13,110
really integrated immediately
into the other

718
00:29:13,110 --> 00:29:13,890
aspects of the face.

719
00:29:13,890 --> 00:29:15,250
They don't stand alone.

720
00:29:15,250 --> 00:29:19,230
They're a part of a context
all interacting together.

721
00:29:19,230 --> 00:29:22,080
Now here's something pretty
striking and interesting.

722
00:29:22,080 --> 00:29:24,430
And we'll talk about some social
implications of this at

723
00:29:24,430 --> 00:29:27,230
the end of the lecture.

724
00:29:27,230 --> 00:29:28,920
You can't ask infants
many things.

725
00:29:28,920 --> 00:29:32,070
But you can record their
behavior and learn some things

726
00:29:32,070 --> 00:29:34,360
about what their minds are
representing and thinking.

727
00:29:34,360 --> 00:29:37,480
And here's how they test an
infant at say six months of

728
00:29:37,480 --> 00:29:41,180
age, can't talk, can't walk,
can't be instructed to push

729
00:29:41,180 --> 00:29:43,130
buttons or fill out surveys.

730
00:29:43,130 --> 00:29:45,720
How do they get inside the mind
of what an infant knows?

731
00:29:45,720 --> 00:29:46,960
What they'll do is they'll
show a face.

732
00:29:46,960 --> 00:29:51,080
And pretend they show this
face all by itself.

733
00:29:51,080 --> 00:29:52,680
Infant looks for a little bit.

734
00:29:52,680 --> 00:29:55,240
Face disappears, and then they
see the face they just saw and

735
00:29:55,240 --> 00:29:56,910
a new face.

736
00:29:56,910 --> 00:29:59,940
And infants who saw this face
a few moments ago will spend

737
00:29:59,940 --> 00:30:04,580
more time looking at the new
face than the old face.

738
00:30:04,580 --> 00:30:06,690
You can track where
their eyes are.

739
00:30:06,690 --> 00:30:09,140
And that's the evidence they
remember the face they saw

740
00:30:09,140 --> 00:30:11,580
before, because they selectively
spend more time on

741
00:30:11,580 --> 00:30:12,830
the new face.

742
00:30:12,830 --> 00:30:15,230
And also they're hungry to
look at the new face.

743
00:30:15,230 --> 00:30:17,350
I've seen this face, been
there, done that.

744
00:30:17,350 --> 00:30:20,310
New face, check out
this person.

745
00:30:20,310 --> 00:30:22,700
If you test them at six
months, they'll have a

746
00:30:22,700 --> 00:30:26,360
preference for the new face.

747
00:30:26,360 --> 00:30:29,810
But if you show them two
different monkeys--

748
00:30:29,810 --> 00:30:32,110
and this might feel hard for
you to tell these apart--

749
00:30:32,110 --> 00:30:34,895
if you show them two different
monkeys, they'll behave the

750
00:30:34,895 --> 00:30:38,440
same with monkey faces as they
do with human faces.

751
00:30:38,440 --> 00:30:41,720
They'll prefer the novel face,
which shows you they remember

752
00:30:41,720 --> 00:30:44,710
what face they saw before and
want to look at the new face.

753
00:30:44,710 --> 00:30:49,330
And at six months, it's the same
for humans and monkeys.

754
00:30:49,330 --> 00:30:51,070
And don't forget, most of these
people are interacting a

755
00:30:51,070 --> 00:30:53,600
lot with adults, and siblings,
and caretakers.

756
00:30:53,600 --> 00:30:57,330
Not many of them are hanging
out with monkeys.

757
00:30:57,330 --> 00:31:01,520
They start to get much more
accurate for human faces.

758
00:31:01,520 --> 00:31:04,420
The experience they have with
human faces starts to make

759
00:31:04,420 --> 00:31:05,920
them much better that.

760
00:31:05,920 --> 00:31:08,490
At birth, faces from
one species

761
00:31:08,490 --> 00:31:09,700
are as good as another.

762
00:31:09,700 --> 00:31:12,240
But with practice and
experience, they become human

763
00:31:12,240 --> 00:31:14,030
face specialists.

764
00:31:14,030 --> 00:31:16,130
And they become much
better at that.

765
00:31:16,130 --> 00:31:19,810
So it's not in our genes to
be awesome at human faces.

766
00:31:19,810 --> 00:31:21,430
But there's something
about faces.

767
00:31:21,430 --> 00:31:24,790
And then a lot depends
on what we see.

768
00:31:24,790 --> 00:31:27,140
And we grow that part of
our mind and brain.

769
00:31:27,140 --> 00:31:29,070
So here's a remarkable
experiment, that you couldn't

770
00:31:29,070 --> 00:31:32,860
do with people, that was done
with infant monkeys.

771
00:31:32,860 --> 00:31:36,030
These monkeys, after they were
born, were given no exposure

772
00:31:36,030 --> 00:31:38,890
to faces, depending on the
monkey, for six to 24 months.

773
00:31:38,890 --> 00:31:41,120
That's not quite completely
true.

774
00:31:41,120 --> 00:31:43,430
Because at their birth,
they saw some faces.

775
00:31:43,430 --> 00:31:45,090
So they couldn't start this
experiment at the

776
00:31:45,090 --> 00:31:45,980
first moment of birth.

777
00:31:45,980 --> 00:31:49,650
But as soon as they could from
birth, for six to 24 months.

778
00:31:49,650 --> 00:31:53,350
So when they were being fed,
this is what they saw.

779
00:31:53,350 --> 00:31:55,010
They were given an interesting
environment

780
00:31:55,010 --> 00:31:56,550
to see lots of stuff.

781
00:31:56,550 --> 00:31:57,730
And they would live in
something like this.

782
00:31:57,730 --> 00:31:59,660
So they could not see
another monkey.

783
00:31:59,660 --> 00:32:00,610
They could not see a human.

784
00:32:00,610 --> 00:32:02,710
They never saw a face
except for the

785
00:32:02,710 --> 00:32:03,960
first moments of birth.

786
00:32:07,640 --> 00:32:11,790
Before they saw any living face
on a monkey or a human,

787
00:32:11,790 --> 00:32:13,680
they would do these kinds of
experiments that you just

788
00:32:13,680 --> 00:32:14,590
heard with infants.

789
00:32:14,590 --> 00:32:18,030
And they would show that these
monkeys would preferentially

790
00:32:18,030 --> 00:32:21,040
look at faces, both human
faces and monkeys.

791
00:32:21,040 --> 00:32:23,110
And they had no preference
between them.

792
00:32:23,110 --> 00:32:25,830
They would always go for a face
compared to an object, a

793
00:32:25,830 --> 00:32:28,120
monkey face or a human face.

794
00:32:28,120 --> 00:32:29,900
And they could do these kinds
of discrimination.

795
00:32:29,900 --> 00:32:32,730
The first moment before they had
ever-- except for whatever

796
00:32:32,730 --> 00:32:35,020
happens right a birth--
seen one.

797
00:32:35,020 --> 00:32:38,640
So they never practiced
on anything.

798
00:32:38,640 --> 00:32:40,540
Because they had not been
allowed to see a face.

799
00:32:40,540 --> 00:32:44,480
But the first moment before they
saw any living face, they

800
00:32:44,480 --> 00:32:48,040
already had this preference for
faces for both humans and

801
00:32:48,040 --> 00:32:52,230
monkeys, as if it's in our genes
to be drawn to faces.

802
00:32:52,230 --> 00:32:55,040
And what the mechanism is, the
top-heavy, something else, we

803
00:32:55,040 --> 00:32:56,040
don't quite know.

804
00:32:56,040 --> 00:32:58,440
But it's in our brain
to go for faces.

805
00:32:58,440 --> 00:33:00,220
They're so important for us.

806
00:33:00,220 --> 00:33:03,410
Then some monkeys hung
around people.

807
00:33:03,410 --> 00:33:06,180
And some monkeys hung
around monkeys.

808
00:33:06,180 --> 00:33:09,650
And what happened is really
quite striking.

809
00:33:09,650 --> 00:33:13,440
They got better at the species
they hung around.

810
00:33:13,440 --> 00:33:16,320
So if you did these kinds of
tests, the monkeys who hung

811
00:33:16,320 --> 00:33:18,970
around monkeys got better
and better at monkeys.

812
00:33:18,970 --> 00:33:20,830
And the monkeys that hung around
people got better and

813
00:33:20,830 --> 00:33:22,740
better at people.

814
00:33:22,740 --> 00:33:24,150
So what does that mean?

815
00:33:24,150 --> 00:33:26,440
That means it looks like there's
a genetic preparation.

816
00:33:26,440 --> 00:33:29,080
The first face that they see
with the slight asterisk of

817
00:33:29,080 --> 00:33:31,700
the moment of birth, they are
already preferring faces.

818
00:33:31,700 --> 00:33:33,800
The very first one they see,
even though they've gone six

819
00:33:33,800 --> 00:33:38,020
months to two years without
seeing any face at all.

820
00:33:38,020 --> 00:33:40,390
And then they get better
selectively on the species

821
00:33:40,390 --> 00:33:41,900
they're exposed to.

822
00:33:41,900 --> 00:33:44,200
So there's a genetic preparation
and then kind of a

823
00:33:44,200 --> 00:33:46,940
sensitive period where you build
up a representation of

824
00:33:46,940 --> 00:33:49,980
the species that you're
living with.

825
00:33:49,980 --> 00:33:54,840
So let me slide there for
a moment to now writing.

826
00:33:54,840 --> 00:33:56,830
As long as there's been mammals,
as long as there's

827
00:33:56,830 --> 00:34:00,535
been people, we've been born to
things around us with faces

828
00:34:00,535 --> 00:34:03,960
and interacted with people
who have faces.

829
00:34:03,960 --> 00:34:05,960
So that's deep in our
evolutionary history.

830
00:34:05,960 --> 00:34:07,890
Not deep in our evolutionary
history

831
00:34:07,890 --> 00:34:09,730
is writing and reading.

832
00:34:09,730 --> 00:34:11,909
The first visual language
of any kind is

833
00:34:11,909 --> 00:34:14,620
about 4,000 years ago.

834
00:34:14,620 --> 00:34:17,409
Not until the Gutenberg
Bible--

835
00:34:17,409 --> 00:34:21,159
and that's about
600 years ago--

836
00:34:21,159 --> 00:34:23,739
did large numbers
of people read.

837
00:34:23,739 --> 00:34:26,860
So unlike faces and objects
which have been around

838
00:34:26,860 --> 00:34:31,510
forever, not only for us, but
for other species, writing and

839
00:34:31,510 --> 00:34:34,080
reading have been around only
for a few hundred years.

840
00:34:34,080 --> 00:34:36,320
Our brains are not evolved
to be reading brains.

841
00:34:36,320 --> 00:34:40,880
It's just an unbelievably
powerful cultural invention.

842
00:34:40,880 --> 00:34:42,610
And people can read fast.

843
00:34:42,610 --> 00:34:44,550
A typical reader can do this.

844
00:34:44,550 --> 00:34:47,199
This is kind of amazing.

845
00:34:47,199 --> 00:34:48,370
Most of us know somewhere
between

846
00:34:48,370 --> 00:34:51,580
50,000 and 100,000 words.

847
00:34:51,580 --> 00:34:52,610
We don't know all
of them well.

848
00:34:52,610 --> 00:34:53,850
We can't spell all of
them correctly.

849
00:34:53,850 --> 00:34:55,690
But we kind of can get
them if we see them.

850
00:34:55,690 --> 00:34:59,750
And if we show you one word in
50,000ths of a second, you can

851
00:34:59,750 --> 00:35:02,210
recognize which word you saw.

852
00:35:02,210 --> 00:35:03,720
It's kind of amazing.

853
00:35:03,720 --> 00:35:06,830
People read about 200 to 250
words per minute, a typical

854
00:35:06,830 --> 00:35:10,000
reader cruising along when
they're focused.

855
00:35:10,000 --> 00:35:11,980
These things can be measured
relatively simply.

856
00:35:11,980 --> 00:35:13,020
Here's one that's a
little trickier.

857
00:35:13,020 --> 00:35:13,560
But let me tell you.

858
00:35:13,560 --> 00:35:15,480
Again, it shows you how
experiments can tell us

859
00:35:15,480 --> 00:35:16,960
something we could not have
guessed without the

860
00:35:16,960 --> 00:35:20,210
experiment, about how you read
every moment you've read

861
00:35:20,210 --> 00:35:21,730
throughout your life.

862
00:35:21,730 --> 00:35:24,450
So I'm going to tell you now
that you read about 12

863
00:35:24,450 --> 00:35:25,470
letters at a time.

864
00:35:25,470 --> 00:35:26,680
And then you move your eyes.

865
00:35:26,680 --> 00:35:29,320
Whatever you're reading, a book,
on the computer, you'll

866
00:35:29,320 --> 00:35:30,970
read about 12 letters.

867
00:35:30,970 --> 00:35:33,300
That's where you're fixated.

868
00:35:33,300 --> 00:35:36,820
Those will only be visible to
you, and then you'll move your

869
00:35:36,820 --> 00:35:40,890
eyes to get the next set of
letters you need to read.

870
00:35:40,890 --> 00:35:42,260
Now you go, no.

871
00:35:42,260 --> 00:35:44,030
There's text there.

872
00:35:44,030 --> 00:35:45,320
I'll give you, I move my eyes.

873
00:35:45,320 --> 00:35:48,660
But what's really happening, let
me show you how blind you

874
00:35:48,660 --> 00:35:52,080
are to everything but 12
letters at a time.

875
00:35:52,080 --> 00:35:54,590
So here's the experiment
from McConkie & Raynor.

876
00:35:54,590 --> 00:35:56,390
And many people have
found this since.

877
00:35:56,390 --> 00:35:58,200
They tracked people's
eye movements.

878
00:35:58,200 --> 00:35:59,620
You can have a machine that
tracks where your eyes are

879
00:35:59,620 --> 00:36:01,890
fixated, where you're looking.

880
00:36:01,890 --> 00:36:04,970
Every time you moved, what they
did is they took the text

881
00:36:04,970 --> 00:36:08,640
that was written across a
monitor, they kept about 12

882
00:36:08,640 --> 00:36:09,900
letters where you're looking.

883
00:36:09,900 --> 00:36:13,690
And they made everything else
on the monitor x's.

884
00:36:13,690 --> 00:36:14,460
You don't know this.

885
00:36:14,460 --> 00:36:16,140
You're just reading.

886
00:36:16,140 --> 00:36:20,460
So you're reading "people of."
This is where your eyes are.

887
00:36:20,460 --> 00:36:22,010
The computer makes everything
else an x.

888
00:36:22,010 --> 00:36:23,870
You move your eyes.

889
00:36:23,870 --> 00:36:24,910
Now this becomes x's.

890
00:36:24,910 --> 00:36:26,030
And these are x's.

891
00:36:26,030 --> 00:36:27,800
You move your eye.

892
00:36:27,800 --> 00:36:28,990
These become x's.

893
00:36:28,990 --> 00:36:30,210
You move your eyes.

894
00:36:30,210 --> 00:36:31,015
These become x's.

895
00:36:31,015 --> 00:36:32,260
Does hat make sense?

896
00:36:32,260 --> 00:36:34,970
Every time you move, where you
land, they give you the words

897
00:36:34,970 --> 00:36:35,740
and letters.

898
00:36:35,740 --> 00:36:38,400
But everything to the left
and right are all x's.

899
00:36:38,400 --> 00:36:42,980
And the astounding thing is
nobody ever says hey, why are

900
00:36:42,980 --> 00:36:45,540
there all the x's.

901
00:36:45,540 --> 00:36:48,570
Because when you're reading,
unbeknownst to you, you only

902
00:36:48,570 --> 00:36:52,300
see with good acuity 12
letters at a time.

903
00:36:52,300 --> 00:36:55,350
And everything else, you barely
notice it's there.

904
00:36:55,350 --> 00:36:57,160
You notice something is there.

905
00:36:57,160 --> 00:36:59,410
If there's nothing there,
you would notice it.

906
00:36:59,410 --> 00:37:03,190
But your mind is not really
seeing it at all.

907
00:37:03,190 --> 00:37:06,700
And so people have tried
to represent this.

908
00:37:06,700 --> 00:37:09,320
This, when you read, is
something like this.

909
00:37:09,320 --> 00:37:11,780
Where you look, in your fovea,
or your high acuity, you see

910
00:37:11,780 --> 00:37:13,030
about 12 letters.

911
00:37:13,030 --> 00:37:16,320
Everything else is like this.

912
00:37:16,320 --> 00:37:19,670
And that's why if I flip it to
x's, you don't even notice.

913
00:37:19,670 --> 00:37:21,730
Because it's all pretty
crummy how you see it.

914
00:37:21,730 --> 00:37:22,885
You notice something is there.

915
00:37:22,885 --> 00:37:26,420
But it's pretty crummy.

916
00:37:26,420 --> 00:37:27,560
You can say, I don't
believe it.

917
00:37:27,560 --> 00:37:28,520
But people have done
this experiment

918
00:37:28,520 --> 00:37:29,530
over and over again.

919
00:37:29,530 --> 00:37:31,650
And when they flip everything
to x's to the left and right

920
00:37:31,650 --> 00:37:33,800
of about 12 letters that you're
looking at, people just

921
00:37:33,800 --> 00:37:34,550
don't know.

922
00:37:34,550 --> 00:37:37,490
Because they don't really see
it as they read anymore.

923
00:37:37,490 --> 00:37:38,510
Does that make sense?

924
00:37:38,510 --> 00:37:40,980
12 letters at a time.

925
00:37:40,980 --> 00:37:43,440
So again, we're trying to
understand some of these

926
00:37:43,440 --> 00:37:46,560
things in regards to
difficulties people have.

927
00:37:46,560 --> 00:37:49,710
So it turns out there's a part
of the brain that's also

928
00:37:49,710 --> 00:37:52,880
involved in reading where, if
you damage, you get what's

929
00:37:52,880 --> 00:37:54,020
called word blindness.

930
00:37:54,020 --> 00:37:55,330
So these are adults
who are good

931
00:37:55,330 --> 00:37:56,860
readers, no problem reading.

932
00:37:56,860 --> 00:37:59,690
They get a stroke Hts left
visual cortex around here.

933
00:37:59,690 --> 00:38:03,470
The most famous original case,
in 1887, all of a sudden one

934
00:38:03,470 --> 00:38:06,020
day after a headache, this
man was unable to read.

935
00:38:06,020 --> 00:38:08,310
His vision was otherwise
perfectly good.

936
00:38:08,310 --> 00:38:10,240
He could hear words
and speak words.

937
00:38:10,240 --> 00:38:12,030
He could see numbers.

938
00:38:12,030 --> 00:38:13,720
He could write down words.

939
00:38:13,720 --> 00:38:15,490
His writing was kind of
slow and laborious.

940
00:38:15,490 --> 00:38:17,230
It wasn't typical writing.

941
00:38:17,230 --> 00:38:19,610
But then you could show him a
few moments later what he had

942
00:38:19,610 --> 00:38:22,570
written, and he would not be
able to read it aloud.

943
00:38:22,570 --> 00:38:26,740
So here's a pathway where stuff
out there in the world

944
00:38:26,740 --> 00:38:28,670
in vision meets how we read.

945
00:38:36,820 --> 00:38:37,200
Amazing.

946
00:38:37,200 --> 00:38:39,920
Now it turns out, just like
there's a spot in your brain

947
00:38:39,920 --> 00:38:43,020
that's the first moment where
something you see is a face.

948
00:38:43,020 --> 00:38:48,240
There's a part that's shown here
in red that's near that.

949
00:38:48,240 --> 00:38:50,150
It's bigger in the left
hemisphere, the language

950
00:38:50,150 --> 00:38:50,950
hemisphere.

951
00:38:50,950 --> 00:38:53,330
It's pretty specific relatively
specific for words

952
00:38:53,330 --> 00:38:56,240
compared to other things.

953
00:38:56,240 --> 00:38:58,690
And so we end up with a
picture like this with

954
00:38:58,690 --> 00:39:02,070
relatively specific areas in the
back of our brain that are

955
00:39:02,070 --> 00:39:06,800
highly specialized for
recognizing where things are.

956
00:39:06,800 --> 00:39:08,020
I didn't talk too
much about that.

957
00:39:08,020 --> 00:39:12,350
But there's that area, faces,
words, and objects.

958
00:39:12,350 --> 00:39:15,300
And if you have damage to any
of these pathways, for

959
00:39:15,300 --> 00:39:17,740
whatever reason, the rest of
our brain can't compensate.

960
00:39:17,740 --> 00:39:20,106
And it's all a big mystery, who
am I looking at, what word

961
00:39:20,106 --> 00:39:23,540
am I looking at, what
object do I see?

962
00:39:23,540 --> 00:39:27,340
Now, you might imagine that
for things like faces or

963
00:39:27,340 --> 00:39:29,880
locations, we've had
evolutionary pressure for

964
00:39:29,880 --> 00:39:31,270
forever to be good at that.

965
00:39:31,270 --> 00:39:32,110
Where are we?

966
00:39:32,110 --> 00:39:33,630
Who are we with?

967
00:39:33,630 --> 00:39:37,960
Whereas things like words, or
nonsense words, and letters,

968
00:39:37,960 --> 00:39:39,080
or chairs--

969
00:39:39,080 --> 00:39:40,950
which are a recent cultural
invention--

970
00:39:40,950 --> 00:39:44,040
our brains are not genetically
predisposed to those specific

971
00:39:44,040 --> 00:39:44,460
categories.

972
00:39:44,460 --> 00:39:47,310
So there's a very clever study
from Ted Polk in Michigan.

973
00:39:47,310 --> 00:39:48,270
So we took twins.

974
00:39:48,270 --> 00:39:49,350
And we'll talk more
about this.

975
00:39:49,350 --> 00:39:52,370
But the way people try to get
at until we know the real

976
00:39:52,370 --> 00:39:55,410
genes involved in things,
they'll compare twins who are

977
00:39:55,410 --> 00:39:58,695
identical twins with the
identical genes and twins

978
00:39:58,695 --> 00:40:02,160
which are so-called fraternal
twins who are born to the same

979
00:40:02,160 --> 00:40:04,870
family but don't have
the identical genes.

980
00:40:04,870 --> 00:40:06,990
And to the extent the identical
twins are more

981
00:40:06,990 --> 00:40:10,290
similar to each other than the
fraternal twins, that's a

982
00:40:10,290 --> 00:40:13,420
suggestion there's genetic
influences on whatever you're

983
00:40:13,420 --> 00:40:14,680
looking at.

984
00:40:14,680 --> 00:40:17,210
So what they did is they had
these people look at faces,

985
00:40:17,210 --> 00:40:20,760
places, letter strings,
and chairs.

986
00:40:20,760 --> 00:40:23,630
And they calculated how similar
was one twin to

987
00:40:23,630 --> 00:40:28,140
another, either identically
genetic or not.

988
00:40:28,140 --> 00:40:32,200
And what they found was that for
the faces and the places,

989
00:40:32,200 --> 00:40:35,100
there was much more similar,
significantly more similarly

990
00:40:35,100 --> 00:40:38,110
between identical twins
than fraternal twins.

991
00:40:38,110 --> 00:40:40,910
That's the genetic
predisposition, in some way,

992
00:40:40,910 --> 00:40:44,600
of how you will see a face
or know where you are.

993
00:40:44,600 --> 00:40:46,640
If they compare reading
nonsense words, letter

994
00:40:46,640 --> 00:40:49,540
strings, or chairs, the
two kinds of twins

995
00:40:49,540 --> 00:40:51,110
were just the same.

996
00:40:51,110 --> 00:40:54,990
It's as if there's no genetic
predisposition in you to know

997
00:40:54,990 --> 00:40:55,960
where in your brain
you're going to

998
00:40:55,960 --> 00:40:57,360
process words or chairs.

999
00:40:57,360 --> 00:40:58,370
But that makes sense.

1000
00:40:58,370 --> 00:41:03,770
Because a couple thousand years
ago there were no words.

1001
00:41:03,770 --> 00:41:05,020
And a few more thousand
years than that,

1002
00:41:05,020 --> 00:41:06,330
there were no chairs.

1003
00:41:06,330 --> 00:41:08,870
So that's we don't have
the genes for recent

1004
00:41:08,870 --> 00:41:10,140
inventions in the world.

1005
00:41:10,140 --> 00:41:12,340
But our brain is brilliant at
becoming good at things like

1006
00:41:12,340 --> 00:41:15,425
reading as we invent them.

1007
00:41:15,425 --> 00:41:17,930
For the last part of the talk,
I want to talk about face

1008
00:41:17,930 --> 00:41:18,470
recognition.

1009
00:41:18,470 --> 00:41:22,090
So all of us know that when
we learn to read it was

1010
00:41:22,090 --> 00:41:22,900
a pretty big deal.

1011
00:41:22,900 --> 00:41:24,300
It was a while ago.

1012
00:41:24,300 --> 00:41:26,360
If you are a typical reader,
you're passionate.

1013
00:41:26,360 --> 00:41:27,550
If you're a struggling
reader, that can be

1014
00:41:27,550 --> 00:41:28,600
a persistent problem.

1015
00:41:28,600 --> 00:41:30,520
But you got through it,
and you're a reader.

1016
00:41:30,520 --> 00:41:33,210
None of us went to third grade
and said, today is the day

1017
00:41:33,210 --> 00:41:35,346
that we study faces.

1018
00:41:35,346 --> 00:41:37,810
And your parents are going, I
hope you're going at faces.

1019
00:41:37,810 --> 00:41:40,390
Because we really want you
to go to MIT someday.

1020
00:41:40,390 --> 00:41:43,270
This is the moment to get
ready for the face SAT.

1021
00:41:43,270 --> 00:41:45,250
But what I'm going to show you
is that in fact, through

1022
00:41:45,250 --> 00:41:48,040
experience, you're educated
for faces.

1023
00:41:48,040 --> 00:41:50,750
And there's some social
consequences for the faces

1024
00:41:50,750 --> 00:41:53,340
you're not educated
on, on average.

1025
00:41:53,340 --> 00:41:55,170
So let's talk about that.

1026
00:41:55,170 --> 00:41:59,210
So one way we see faces is we
see faces that are right-side

1027
00:41:59,210 --> 00:42:01,580
up, overwhelmingly.

1028
00:42:01,580 --> 00:42:04,210
So I'm going to show you
this upside down face.

1029
00:42:04,210 --> 00:42:05,380
I'm going to reverse it.

1030
00:42:05,380 --> 00:42:07,070
And that's what it is.

1031
00:42:07,070 --> 00:42:08,840
That's gruesome, right?

1032
00:42:08,840 --> 00:42:11,220
But can you tell that
when that turns,

1033
00:42:11,220 --> 00:42:12,800
it's going to be this?

1034
00:42:12,800 --> 00:42:15,930
So the way that we interpret
the weirdness of face

1035
00:42:15,930 --> 00:42:18,780
inversion, the fact that
we process upside

1036
00:42:18,780 --> 00:42:22,030
down faces so badly--

1037
00:42:22,030 --> 00:42:24,230
we don't have any hint of
what's really there--

1038
00:42:24,230 --> 00:42:26,800
it's because we didn't
practice on them.

1039
00:42:26,800 --> 00:42:29,310
It's as if I gave you a test in
a language that you never

1040
00:42:29,310 --> 00:42:31,030
spoke or read.

1041
00:42:31,030 --> 00:42:32,920
Well you can do badly.

1042
00:42:32,920 --> 00:42:34,845
So that shows you that
faces, it's a

1043
00:42:34,845 --> 00:42:36,510
so experienced dependent.

1044
00:42:36,510 --> 00:42:38,520
It's because you see upright
faces you get

1045
00:42:38,520 --> 00:42:39,470
good at upright faces.

1046
00:42:39,470 --> 00:42:41,000
All we have to do is turn
it upside down.

1047
00:42:41,000 --> 00:42:44,230
And all the knowledge and skill
you have, it's as if I

1048
00:42:44,230 --> 00:42:45,920
tested you in a different
language.

1049
00:42:45,920 --> 00:42:48,530
It's that specific.

1050
00:42:48,530 --> 00:42:51,450
Here's another example, side by
side, looks pretty similar.

1051
00:42:51,450 --> 00:42:52,700
But that's what you're
really seeing.

1052
00:42:57,390 --> 00:42:59,220
It's unbelievable.

1053
00:42:59,220 --> 00:43:03,400
Turn your head to believe me
in case you're suspicious.

1054
00:43:03,400 --> 00:43:04,650
This is legit.

1055
00:43:06,560 --> 00:43:09,830
We're so bad at upside down
faces because we don't

1056
00:43:09,830 --> 00:43:12,270
practice on them.

1057
00:43:12,270 --> 00:43:15,640
So what we are good at in faces
is so much dependent on

1058
00:43:15,640 --> 00:43:17,930
what we practice.

1059
00:43:17,930 --> 00:43:19,240
You may recognize this
person or not.

1060
00:43:24,440 --> 00:43:25,690
These are less dramatic.

1061
00:43:28,930 --> 00:43:30,670
People don't become as good
as they get at faces

1062
00:43:30,670 --> 00:43:32,870
until about age 16.

1063
00:43:32,870 --> 00:43:34,680
We don't consciously
practice it.

1064
00:43:34,680 --> 00:43:36,890
But the social experience
we have is practicing,

1065
00:43:36,890 --> 00:43:40,240
practicing, practicing how
you perceive a face.

1066
00:43:43,670 --> 00:43:46,530
One fun study from a Susan Carey
now at Harvard is she

1067
00:43:46,530 --> 00:43:48,580
took dog show judges.

1068
00:43:48,580 --> 00:43:50,300
Now, I don't know too much
about dog shows.

1069
00:43:50,300 --> 00:43:51,850
Some of you might
know something.

1070
00:43:51,850 --> 00:43:53,510
But a little bit of what
I know is this.

1071
00:43:53,510 --> 00:43:56,670
Sort of like in sports, you have
judges who work their way

1072
00:43:56,670 --> 00:44:00,260
up a ladder of experience
and reputation.

1073
00:44:00,260 --> 00:44:02,620
So in sports things, you
have the local judges.

1074
00:44:02,620 --> 00:44:03,940
And the really good ones
become the ones

1075
00:44:03,940 --> 00:44:04,810
for the bigger leagues.

1076
00:44:04,810 --> 00:44:06,660
And the big ones get to go to
the Super Bowl and World

1077
00:44:06,660 --> 00:44:09,620
Series who are really skilled
and thought to be excellent.

1078
00:44:09,620 --> 00:44:11,120
Dog show judges, similarly,
start locally.

1079
00:44:11,120 --> 00:44:13,420
The ones who seem pretty good
and persistent at it then

1080
00:44:13,420 --> 00:44:14,480
become the state judge show.

1081
00:44:14,480 --> 00:44:16,940
And then they go to the
big one in New York.

1082
00:44:16,940 --> 00:44:21,870
And so they did these face
inversion experiments on dog

1083
00:44:21,870 --> 00:44:24,970
show judges, people who were
just beginning to be a dog

1084
00:44:24,970 --> 00:44:27,300
show judge, people who have done
it for a while, people

1085
00:44:27,300 --> 00:44:28,790
who've done it for
a lot of years.

1086
00:44:28,790 --> 00:44:32,320
And they asked, when do you
get so much better at a

1087
00:44:32,320 --> 00:44:35,270
right-side up dog face than
an upside down dog face?

1088
00:44:35,270 --> 00:44:38,350
Now, most people are about
the same to start with.

1089
00:44:38,350 --> 00:44:40,830
Because we're not that
good at dog faces.

1090
00:44:40,830 --> 00:44:41,990
You might know a dog or two.

1091
00:44:41,990 --> 00:44:45,060
But you don't know hundreds
of dollars.

1092
00:44:45,060 --> 00:44:48,660
It took eight years for dog show
judges-- and they were

1093
00:44:48,660 --> 00:44:50,910
dog enthusiasts for many
years before they

1094
00:44:50,910 --> 00:44:52,490
became a dog show judge--

1095
00:44:52,490 --> 00:44:56,480
eight years of seeing many,
many dogs at dog show

1096
00:44:56,480 --> 00:45:00,600
competitions to develop the
upright face expertise that we

1097
00:45:00,600 --> 00:45:02,720
all have for faces as adults.

1098
00:45:02,720 --> 00:45:05,750
So for an adult, you have to
practice eight years to get

1099
00:45:05,750 --> 00:45:08,140
this kind of expertise.

1100
00:45:08,140 --> 00:45:09,930
You can't just do
it by cramming.

1101
00:45:09,930 --> 00:45:12,430
It's practice, practice,
practice in the most brutal

1102
00:45:12,430 --> 00:45:13,680
fundamental sense.

1103
00:45:15,810 --> 00:45:18,410
Now I'll say one last thing
about this then

1104
00:45:18,410 --> 00:45:19,850
go to the last topic.

1105
00:45:19,850 --> 00:45:21,940
The other thing is that we
understand also-- and we've

1106
00:45:21,940 --> 00:45:24,312
mentioned this before, but I'll
say it one more time--

1107
00:45:24,312 --> 00:45:28,700
that imagination and perception
use the same neural

1108
00:45:28,700 --> 00:45:30,170
territory in your mind.

1109
00:45:30,170 --> 00:45:32,610
So here's a study from Nancy
Kanwisher again.

1110
00:45:32,610 --> 00:45:35,760
Here's a part of the brain that
responds to faces, seeing

1111
00:45:35,760 --> 00:45:39,600
a face, seeing a house, seeing
a face, seeing a house.

1112
00:45:39,600 --> 00:45:43,380
Now, imagining a face, imagining
a house, imagining a

1113
00:45:43,380 --> 00:45:45,720
face, imagining a house.

1114
00:45:45,720 --> 00:45:49,460
So seeing one drives the
brain more strongly

1115
00:45:49,460 --> 00:45:50,910
than imagining it.

1116
00:45:50,910 --> 00:45:54,030
But when you imagine a face, you
seem to use the same brain

1117
00:45:54,030 --> 00:45:56,600
network as when you see it.

1118
00:45:56,600 --> 00:45:59,230
It's just not as vivid
and powerful.

1119
00:45:59,230 --> 00:46:02,280
So perception is the stuff
literally in your mind of

1120
00:46:02,280 --> 00:46:03,000
imagination.

1121
00:46:03,000 --> 00:46:05,870
You can imagine things, because
you've seen them

1122
00:46:05,870 --> 00:46:08,450
before, literally in
your brain It's the

1123
00:46:08,450 --> 00:46:11,090
same part of the brain.

1124
00:46:11,090 --> 00:46:13,620
So the last topic I want to talk
about is a sensitive and

1125
00:46:13,620 --> 00:46:14,370
difficult one.

1126
00:46:14,370 --> 00:46:17,220
So I'm going to do
it right, I hope.

1127
00:46:17,220 --> 00:46:19,250
But it's one that's a
practical one, and a

1128
00:46:19,250 --> 00:46:23,240
consequence, we believe, of the
fact that you're good at

1129
00:46:23,240 --> 00:46:25,120
faces that you practice.

1130
00:46:25,120 --> 00:46:28,600
So we're going to talk about
in these studies people

1131
00:46:28,600 --> 00:46:31,950
perceiving faces of European
Americans, or white Americans,

1132
00:46:31,950 --> 00:46:33,940
or African Americans.

1133
00:46:33,940 --> 00:46:37,630
And there's a huge amount
of research

1134
00:46:37,630 --> 00:46:39,120
that shows the following.

1135
00:46:39,120 --> 00:46:44,920
That whatever racial group you
see a lot of, you're pretty

1136
00:46:44,920 --> 00:46:46,240
good at remembering
those faces.

1137
00:46:46,240 --> 00:46:50,790
At whatever group you see
less, you're far worse.

1138
00:46:50,790 --> 00:46:53,950
We interpret this as a
consequence of the practice

1139
00:46:53,950 --> 00:46:55,720
that you have.

1140
00:46:55,720 --> 00:46:58,900
But social psychologists worry
it has a second consequence.

1141
00:46:58,900 --> 00:47:01,150
Let me tell you the
practice story.

1142
00:47:01,150 --> 00:47:02,540
So we did a study.

1143
00:47:02,540 --> 00:47:03,120
This is from us.

1144
00:47:03,120 --> 00:47:04,800
But it's very similar to other
people in terms of the

1145
00:47:04,800 --> 00:47:05,500
conclusion.

1146
00:47:05,500 --> 00:47:08,500
There's literally like 1,000
of these studies that show

1147
00:47:08,500 --> 00:47:11,480
that people are superior at
recognizing in an experiment

1148
00:47:11,480 --> 00:47:14,950
faces from their own group
than from other groups.

1149
00:47:14,950 --> 00:47:17,300
And here, an experiment that
was done with Stanford

1150
00:47:17,300 --> 00:47:19,850
students, a look in their mind
and then in their brain.

1151
00:47:19,850 --> 00:47:21,930
So these are Stanford students
just some years ago, maybe a

1152
00:47:21,930 --> 00:47:23,180
lot like you in many ways.

1153
00:47:25,820 --> 00:47:28,550
They saw some faces, and then
they got a memory test.

1154
00:47:28,550 --> 00:47:30,790
This is their performance
on the memory test.

1155
00:47:30,790 --> 00:47:35,050
So we had both European American
and African American

1156
00:47:35,050 --> 00:47:37,650
faces that they viewed.

1157
00:47:37,650 --> 00:47:40,410
And the participants in the
research were both African

1158
00:47:40,410 --> 00:47:42,470
American and European
American.

1159
00:47:42,470 --> 00:47:46,370
OK so there were white faces,
white subjects, black faces,

1160
00:47:46,370 --> 00:47:48,070
black subjects.

1161
00:47:48,070 --> 00:47:49,800
And what you can see is this.

1162
00:47:49,800 --> 00:47:51,820
The higher the bar, the
better the memory.

1163
00:47:51,820 --> 00:47:55,440
In both groups, they're better
in performance at remembering

1164
00:47:55,440 --> 00:48:00,420
faces from their own group.

1165
00:48:00,420 --> 00:48:02,350
But there's more to
it than that.

1166
00:48:02,350 --> 00:48:05,590
If you look at these bars, these
are the memory for the

1167
00:48:05,590 --> 00:48:06,695
white faces.

1168
00:48:06,695 --> 00:48:09,240
And it's about the same
in the two groups.

1169
00:48:09,240 --> 00:48:11,120
And look at the big
difference for the

1170
00:48:11,120 --> 00:48:13,190
African American faces.

1171
00:48:13,190 --> 00:48:15,360
So two things are
happening here.

1172
00:48:15,360 --> 00:48:18,610
One of them is there's superior
memory from faces

1173
00:48:18,610 --> 00:48:20,630
from your own group.

1174
00:48:20,630 --> 00:48:25,480
And the second story is that for
what we believed, and from

1175
00:48:25,480 --> 00:48:27,740
other research as
well, is this.

1176
00:48:27,740 --> 00:48:30,670
If you're a minority group,
you're still pretty good at

1177
00:48:30,670 --> 00:48:32,420
the majority faces.

1178
00:48:32,420 --> 00:48:34,460
If you're a majority group,
you're not so good at the

1179
00:48:34,460 --> 00:48:35,790
minority faces.

1180
00:48:35,790 --> 00:48:38,580
So these are European American
Stanford students.

1181
00:48:38,580 --> 00:48:41,990
And look how a miserable their
memory is for these African

1182
00:48:41,990 --> 00:48:43,240
American faces.

1183
00:48:46,140 --> 00:48:48,920
This is a difficult topic for
us, in part, because race is a

1184
00:48:48,920 --> 00:48:51,980
different topic for
good reasons.

1185
00:48:51,980 --> 00:48:56,560
But social psychologists worry
that the lack of practice that

1186
00:48:56,560 --> 00:49:00,170
many of us have on other groups
growing up at home, in

1187
00:49:00,170 --> 00:49:01,370
our community, and schools--

1188
00:49:01,370 --> 00:49:03,410
and it varies for people,
but on average--

1189
00:49:03,410 --> 00:49:06,710
makes us less able to
individuate in our memory one

1190
00:49:06,710 --> 00:49:09,420
person from another
in other groups.

1191
00:49:09,420 --> 00:49:13,460
And now this is not about
racial attitudes.

1192
00:49:13,460 --> 00:49:17,280
This is simple, objective
recognition of faces.

1193
00:49:17,280 --> 00:49:20,880
But it could certainly support
consideration of other groups

1194
00:49:20,880 --> 00:49:22,670
as if they were all the same.

1195
00:49:22,670 --> 00:49:24,490
Because they look
all the same.

1196
00:49:24,490 --> 00:49:26,430
I remember I had an
African American

1197
00:49:26,430 --> 00:49:28,690
colleague at Stanford.

1198
00:49:28,690 --> 00:49:29,680
And he was a dean.

1199
00:49:29,680 --> 00:49:33,190
And he went to China with a
Nobel laureate who is now the

1200
00:49:33,190 --> 00:49:37,140
Secretary of Energy
in the Obama

1201
00:49:37,140 --> 00:49:39,230
administration, Steve Chu.

1202
00:49:39,230 --> 00:49:41,980
And, of course, Nobel laureates
are a pretty big.

1203
00:49:41,980 --> 00:49:45,540
So this dean went to this Nobel
laureate all the time.

1204
00:49:45,540 --> 00:49:48,610
We're meeting Stanford donors
and other things.

1205
00:49:48,610 --> 00:49:49,710
He said they got on the plane.

1206
00:49:49,710 --> 00:49:52,570
They went to China,
the two of them.

1207
00:49:52,570 --> 00:49:54,390
And he went into a reception.

1208
00:49:54,390 --> 00:49:57,870
And he knew Steve Chu
was in there.

1209
00:49:57,870 --> 00:50:00,500
The African American dean said,
I could not pick him out

1210
00:50:00,500 --> 00:50:01,700
for my life.

1211
00:50:01,700 --> 00:50:05,540
When I saw him now, not with
other people I know, but with

1212
00:50:05,540 --> 00:50:08,790
a 500 other Chinese people
in a large reception.

1213
00:50:08,790 --> 00:50:12,180
So it's not about bad attitudes
in this case.

1214
00:50:12,180 --> 00:50:14,610
It's just literally the
experience you have.

1215
00:50:14,610 --> 00:50:19,960
And in the brain, this part of
this FFA, responds more to

1216
00:50:19,960 --> 00:50:21,600
groups from your own--

1217
00:50:21,600 --> 00:50:22,350
on average--

1218
00:50:22,350 --> 00:50:24,390
from your own racial group than
the other racial group.

1219
00:50:24,390 --> 00:50:26,600
And we think that's promoting
the ease with which to

1220
00:50:26,600 --> 00:50:31,130
identify individuals in your
group and the difficulty with

1221
00:50:31,130 --> 00:50:34,340
which you can identify I
individuals in another group.

1222
00:50:34,340 --> 00:50:36,650
And so we have strong reason
to think that this is all

1223
00:50:36,650 --> 00:50:38,910
about this particular phenomenon
about your

1224
00:50:38,910 --> 00:50:40,340
experience.

1225
00:50:40,340 --> 00:50:41,810
So let me tell you why.

1226
00:50:41,810 --> 00:50:44,400
One thing is we know that this
is not present at birth.

1227
00:50:44,400 --> 00:50:47,040
People have done these kinds of
experiments with infants.

1228
00:50:47,040 --> 00:50:51,300
These racial biases in what
kinds of faces we can remember

1229
00:50:51,300 --> 00:50:52,390
are not present at birth.

1230
00:50:52,390 --> 00:50:54,100
So they're not in our genes.

1231
00:50:54,100 --> 00:50:55,850
They're from our experience.

1232
00:50:55,850 --> 00:50:57,510
But they are present
by three months.

1233
00:50:57,510 --> 00:51:01,180
But by then you've seen a lot
of people in your family.

1234
00:51:01,180 --> 00:51:03,840
And the kind of experiment
that's very telling for this

1235
00:51:03,840 --> 00:51:05,980
is a study where there
were Korean children.

1236
00:51:05,980 --> 00:51:07,230
They were three to
nine years old.

1237
00:51:07,230 --> 00:51:10,850
They were adopted by European
Caucasian families, so, in the

1238
00:51:10,850 --> 00:51:12,950
sense, a cross-racial
adoption.

1239
00:51:12,950 --> 00:51:15,290
And then they tested their
memory for faces.

1240
00:51:15,290 --> 00:51:17,960
And for these Korean children
adopted by European Caucasian

1241
00:51:17,960 --> 00:51:20,900
families, they had better
memory for the Caucasian

1242
00:51:20,900 --> 00:51:26,370
faces, because that's who they
saw around them all the time.

1243
00:51:26,370 --> 00:51:31,810
And they were just like their
French peers who also saw

1244
00:51:31,810 --> 00:51:33,530
Caucasian faces all the time.

1245
00:51:33,530 --> 00:51:35,680
And the opposite of the findings
that you've got with

1246
00:51:35,680 --> 00:51:39,780
Korean children in
South Korea.

1247
00:51:39,780 --> 00:51:43,820
The inversion one is kind
of funny, amazing.

1248
00:51:43,820 --> 00:51:48,250
The fact that we can struggle
to individuate in memory one

1249
00:51:48,250 --> 00:51:51,130
African American from another,
or one Asian American from

1250
00:51:51,130 --> 00:51:53,530
another, or one Caucasian
American from another

1251
00:51:53,530 --> 00:51:55,970
depending on the environment
that we grew up in and were

1252
00:51:55,970 --> 00:52:01,180
exposed to, that could be
facilitating stereotyping and

1253
00:52:01,180 --> 00:52:04,750
lack of individuation of
people in other groups.

1254
00:52:04,750 --> 00:52:06,160
So the very last thing I want
to do-- and this is just a

1255
00:52:06,160 --> 00:52:06,910
couple minutes--

1256
00:52:06,910 --> 00:52:09,270
is talk about expression
in faces.

1257
00:52:09,270 --> 00:52:10,150
And I'll come back to
this in detail.

1258
00:52:10,150 --> 00:52:16,730
I just want to give you one last
example of recognition in

1259
00:52:16,730 --> 00:52:18,130
the amygdala and fear.

1260
00:52:18,130 --> 00:52:20,050
So we'll come back to this
in some detail later on.

1261
00:52:20,050 --> 00:52:23,680
But most researchers argue
there's about six basic

1262
00:52:23,680 --> 00:52:27,040
emotions that people around the
world transmit with their

1263
00:52:27,040 --> 00:52:32,060
facial expressions, happiness,
sadness, fear, anger,

1264
00:52:32,060 --> 00:52:33,620
surprise, and disgust.

1265
00:52:33,620 --> 00:52:36,470
You can show these pictures to
people around the world, and

1266
00:52:36,470 --> 00:52:39,610
they'll be, to some degree,
consensus about what they are,

1267
00:52:39,610 --> 00:52:43,220
universal expressions of
feelings transmitted by facial

1268
00:52:43,220 --> 00:52:45,240
expressions.

1269
00:52:45,240 --> 00:52:46,950
Now let's talk about
fear faces.

1270
00:52:46,950 --> 00:52:48,310
I'll just talk about
fear today.

1271
00:52:48,310 --> 00:52:49,560
And we'll talk about
the amygdala.

1272
00:52:49,560 --> 00:52:51,280
So we have one on the left and
one on the right, it's a very

1273
00:52:51,280 --> 00:52:54,730
small structure, almond shaped,
one inside each of the

1274
00:52:54,730 --> 00:52:58,050
temporal lobes right next
to the hippocampus.

1275
00:52:58,050 --> 00:53:00,610
So that's a view from the side
of where that's located, one

1276
00:53:00,610 --> 00:53:02,540
on the left and one
on the right.

1277
00:53:02,540 --> 00:53:04,360
And we know from a lot of animal
research-- and I'll

1278
00:53:04,360 --> 00:53:06,490
come back to this again--

1279
00:53:06,490 --> 00:53:09,270
the amygdala is essential
for learned

1280
00:53:09,270 --> 00:53:11,590
fear, for learned fear.

1281
00:53:11,590 --> 00:53:14,520
So here's an experiment
that shows you that.

1282
00:53:14,520 --> 00:53:17,430
One thing that mice know
is cats are bad news.

1283
00:53:17,430 --> 00:53:18,930
It's survival.

1284
00:53:18,930 --> 00:53:20,290
What they did in this
experiment is they

1285
00:53:20,290 --> 00:53:21,660
anesthetized the cat.

1286
00:53:21,660 --> 00:53:24,690
That's why the cat looks
glassy-eyed.

1287
00:53:24,690 --> 00:53:29,220
Obviously the experiment would
end too quickly for the mouse,

1288
00:53:29,220 --> 00:53:32,200
if the mouse had to come out
and play with the cat.

1289
00:53:32,200 --> 00:53:33,660
The experiment would be
brutal, and ugly,

1290
00:53:33,660 --> 00:53:35,360
and over too quickly.

1291
00:53:35,360 --> 00:53:38,150
But if the cat is breathing, the
cat's eyes are open, it's

1292
00:53:38,150 --> 00:53:38,510
anesthetized.

1293
00:53:38,510 --> 00:53:40,070
But it can't move.

1294
00:53:40,070 --> 00:53:44,550
Now a typical rat or mouse
will typically

1295
00:53:44,550 --> 00:53:45,670
still avoid the cat.

1296
00:53:45,670 --> 00:53:47,110
It's not worth figuring
out exactly what's

1297
00:53:47,110 --> 00:53:47,910
going on with the cat.

1298
00:53:47,910 --> 00:53:48,970
Are they faking me out?

1299
00:53:48,970 --> 00:53:51,620
I'll just go that way.

1300
00:53:51,620 --> 00:53:55,775
But this is not the mouse who
has the amygdala removed.

1301
00:53:58,500 --> 00:54:00,980
It's not fearful of what it
ought to be fearful of

1302
00:54:00,980 --> 00:54:04,260
anymore, even survival.

1303
00:54:04,260 --> 00:54:07,670
What we believe the most
important role of fear is not

1304
00:54:07,670 --> 00:54:09,530
fear of SATs or midterms.

1305
00:54:09,530 --> 00:54:11,600
Although we live
in that world.

1306
00:54:11,600 --> 00:54:13,950
It's fear of what injure
you and kill you.

1307
00:54:13,950 --> 00:54:14,960
Those are the things to learn.

1308
00:54:14,960 --> 00:54:16,940
Be fearful of what
can kill you.

1309
00:54:16,940 --> 00:54:18,740
Be fearful of what
can injure you.

1310
00:54:18,740 --> 00:54:21,490
Without the amygdala,
you lose that fear.

1311
00:54:21,490 --> 00:54:24,390
But here's the other-- this is
my last slide-- amazing thing.

1312
00:54:24,390 --> 00:54:26,610
There are humans who have
injuries selectively to the

1313
00:54:26,610 --> 00:54:27,620
amygdala, very few.

1314
00:54:27,620 --> 00:54:29,190
But there's a few.

1315
00:54:29,190 --> 00:54:32,550
And not only do they have some
problems with fear learning.

1316
00:54:32,550 --> 00:54:35,640
But they can't recognize
fearful

1317
00:54:35,640 --> 00:54:37,195
facial expressions anymore.

1318
00:54:39,970 --> 00:54:42,350
If they ask them to say, what
is this, without the word

1319
00:54:42,350 --> 00:54:44,440
present, they're very
bad at telling you

1320
00:54:44,440 --> 00:54:45,820
that's a fearful face.

1321
00:54:45,820 --> 00:54:48,180
They can recognize all the other
expressions pretty well

1322
00:54:48,180 --> 00:54:51,680
but not the fearful ones.

1323
00:54:51,680 --> 00:54:53,160
also

1324
00:54:53,160 --> 00:54:54,480
They're good of faces.

1325
00:54:54,480 --> 00:54:55,360
They're not prosopagnosia.

1326
00:54:55,360 --> 00:54:56,160
They're fine at faces.

1327
00:54:56,160 --> 00:54:57,030
They recognize their family.

1328
00:54:57,030 --> 00:54:59,360
But they can't recognize
fearful expressions

1329
00:54:59,360 --> 00:55:01,970
selectively.

1330
00:55:01,970 --> 00:55:05,440
This is actually a very good
drawing happy, sad, surprised,

1331
00:55:05,440 --> 00:55:07,400
disgusted, angry.

1332
00:55:07,400 --> 00:55:09,875
When she's asked to draw a
fearful facial expression,

1333
00:55:09,875 --> 00:55:12,140
this patient draws this,
a baby being afraid.

1334
00:55:12,140 --> 00:55:14,210
She understands the concept.

1335
00:55:14,210 --> 00:55:16,430
But it's not only that she
can't recognize it.

1336
00:55:16,430 --> 00:55:20,020
She can't imagine what a fearful
face looks like.

1337
00:55:20,020 --> 00:55:23,410
And we saw this over
and over again.

1338
00:55:23,410 --> 00:55:25,060
And there's one last
message in this.

1339
00:55:25,060 --> 00:55:27,850
Somehow our knowledge about
what a fearful face looks

1340
00:55:27,850 --> 00:55:29,550
like, we're not scared
of this.

1341
00:55:29,550 --> 00:55:31,130
You don't look at this
and go, oh my gosh.

1342
00:55:31,130 --> 00:55:33,030
I can't believe he showed
that in class.

1343
00:55:33,030 --> 00:55:37,250
I'll never sleep for 48 hours.

1344
00:55:37,250 --> 00:55:39,530
That's a social signal for
somebody telling you, whoa.

1345
00:55:39,530 --> 00:55:41,000
There's something terrible
happening here.

1346
00:55:41,000 --> 00:55:41,670
Look around you.

1347
00:55:41,670 --> 00:55:42,200
We better run.

1348
00:55:42,200 --> 00:55:43,550
It's a social signal.

1349
00:55:43,550 --> 00:55:46,780
And the part of your brain that
learns what can kill you

1350
00:55:46,780 --> 00:55:51,160
also is where the knowledge is
stored to recognize a social

1351
00:55:51,160 --> 00:55:54,160
signal about something that
could be very dangerous.

1352
00:55:54,160 --> 00:55:56,560
The same idea that we put our
knowledge in the parts of

1353
00:55:56,560 --> 00:55:58,180
their brain that act
upon the world.