TEDGlobal 2010
Laurie Santos: A monkey economy as irrational as ours
Filmed:
Readability: 3.3
1,506,660 views
Laurie Santos looks for the roots of human irrationality by watching the way our primate relatives make decisions. A clever series of experiments in "monkeynomics" shows that some of the silly choices we make, monkeys make too.
Laurie Santos - Cognitive psychologist
Laurie Santos studies primate psychology and monkeynomics -- testing problems in human psychology on primates, who (not so surprisingly) have many of the same predictable irrationalities we do. Full bio
Laurie Santos studies primate psychology and monkeynomics -- testing problems in human psychology on primates, who (not so surprisingly) have many of the same predictable irrationalities we do. Full bio
Double-click the English transcript below to play the video.
00:17
I want to start my talk today with two observations
0
2000
2000
00:19
about the human species.
1
4000
2000
00:21
The first observation is something that you might think is quite obvious,
2
6000
3000
00:24
and that's that our species, Homo sapiens,
3
9000
2000
00:26
is actually really, really smart --
4
11000
2000
00:28
like, ridiculously smart --
5
13000
2000
00:30
like you're all doing things
6
15000
2000
00:32
that no other species on the planet does right now.
7
17000
3000
00:35
And this is, of course,
8
20000
2000
00:37
not the first time you've probably recognized this.
9
22000
2000
00:39
Of course, in addition to being smart, we're also an extremely vain species.
10
24000
3000
00:42
So we like pointing out the fact that we're smart.
11
27000
3000
00:45
You know, so I could turn to pretty much any sage
12
30000
2000
00:47
from Shakespeare to Stephen Colbert
13
32000
2000
00:49
to point out things like the fact that
14
34000
2000
00:51
we're noble in reason and infinite in faculties
15
36000
2000
00:53
and just kind of awesome-er than anything else on the planet
16
38000
2000
00:55
when it comes to all things cerebral.
17
40000
3000
00:58
But of course, there's a second observation about the human species
18
43000
2000
01:00
that I want to focus on a little bit more,
19
45000
2000
01:02
and that's the fact that
20
47000
2000
01:04
even though we're actually really smart, sometimes uniquely smart,
21
49000
3000
01:07
we can also be incredibly, incredibly dumb
22
52000
3000
01:10
when it comes to some aspects of our decision making.
23
55000
3000
01:13
Now I'm seeing lots of smirks out there.
24
58000
2000
01:15
Don't worry, I'm not going to call anyone in particular out
25
60000
2000
01:17
on any aspects of your own mistakes.
26
62000
2000
01:19
But of course, just in the last two years
27
64000
2000
01:21
we see these unprecedented examples of human ineptitude.
28
66000
3000
01:24
And we've watched as the tools we uniquely make
29
69000
3000
01:27
to pull the resources out of our environment
30
72000
2000
01:29
kind of just blow up in our face.
31
74000
2000
01:31
We've watched the financial markets that we uniquely create --
32
76000
2000
01:33
these markets that were supposed to be foolproof --
33
78000
3000
01:36
we've watched them kind of collapse before our eyes.
34
81000
2000
01:38
But both of these two embarrassing examples, I think,
35
83000
2000
01:40
don't highlight what I think is most embarrassing
36
85000
3000
01:43
about the mistakes that humans make,
37
88000
2000
01:45
which is that we'd like to think that the mistakes we make
38
90000
3000
01:48
are really just the result of a couple bad apples
39
93000
2000
01:50
or a couple really sort of FAIL Blog-worthy decisions.
40
95000
3000
01:53
But it turns out, what social scientists are actually learning
41
98000
3000
01:56
is that most of us, when put in certain contexts,
42
101000
3000
01:59
will actually make very specific mistakes.
43
104000
3000
02:02
The errors we make are actually predictable.
44
107000
2000
02:04
We make them again and again.
45
109000
2000
02:06
And they're actually immune to lots of evidence.
46
111000
2000
02:08
When we get negative feedback,
47
113000
2000
02:10
we still, the next time we're face with a certain context,
48
115000
3000
02:13
tend to make the same errors.
49
118000
2000
02:15
And so this has been a real puzzle to me
50
120000
2000
02:17
as a sort of scholar of human nature.
51
122000
2000
02:19
What I'm most curious about is,
52
124000
2000
02:21
how is a species that's as smart as we are
53
126000
3000
02:24
capable of such bad
54
129000
2000
02:26
and such consistent errors all the time?
55
131000
2000
02:28
You know, we're the smartest thing out there, why can't we figure this out?
56
133000
3000
02:31
In some sense, where do our mistakes really come from?
57
136000
3000
02:34
And having thought about this a little bit, I see a couple different possibilities.
58
139000
3000
02:37
One possibility is, in some sense, it's not really our fault.
59
142000
3000
02:40
Because we're a smart species,
60
145000
2000
02:42
we can actually create all kinds of environments
61
147000
2000
02:44
that are super, super complicated,
62
149000
2000
02:46
sometimes too complicated for us to even actually understand,
63
151000
3000
02:49
even though we've actually created them.
64
154000
2000
02:51
We create financial markets that are super complex.
65
156000
2000
02:53
We create mortgage terms that we can't actually deal with.
66
158000
3000
02:56
And of course, if we are put in environments where we can't deal with it,
67
161000
3000
02:59
in some sense makes sense that we actually
68
164000
2000
03:01
might mess certain things up.
69
166000
2000
03:03
If this was the case, we'd have a really easy solution
70
168000
2000
03:05
to the problem of human error.
71
170000
2000
03:07
We'd actually just say, okay, let's figure out
72
172000
2000
03:09
the kinds of technologies we can't deal with,
73
174000
2000
03:11
the kinds of environments that are bad --
74
176000
2000
03:13
get rid of those, design things better,
75
178000
2000
03:15
and we should be the noble species
76
180000
2000
03:17
that we expect ourselves to be.
77
182000
2000
03:19
But there's another possibility that I find a little bit more worrying,
78
184000
3000
03:22
which is, maybe it's not our environments that are messed up.
79
187000
3000
03:25
Maybe it's actually us that's designed badly.
80
190000
3000
03:28
This is a hint that I've gotten
81
193000
2000
03:30
from watching the ways that social scientists have learned about human errors.
82
195000
3000
03:33
And what we see is that people tend to keep making errors
83
198000
3000
03:36
exactly the same way, over and over again.
84
201000
3000
03:39
It feels like we might almost just be built
85
204000
2000
03:41
to make errors in certain ways.
86
206000
2000
03:43
This is a possibility that I worry a little bit more about,
87
208000
3000
03:46
because, if it's us that's messed up,
88
211000
2000
03:48
it's not actually clear how we go about dealing with it.
89
213000
2000
03:50
We might just have to accept the fact that we're error prone
90
215000
3000
03:53
and try to design things around it.
91
218000
2000
03:55
So this is the question my students and I wanted to get at.
92
220000
3000
03:58
How can we tell the difference between possibility one and possibility two?
93
223000
3000
04:01
What we need is a population
94
226000
2000
04:03
that's basically smart, can make lots of decisions,
95
228000
2000
04:05
but doesn't have access to any of the systems we have,
96
230000
2000
04:07
any of the things that might mess us up --
97
232000
2000
04:09
no human technology, human culture,
98
234000
2000
04:11
maybe even not human language.
99
236000
2000
04:13
And so this is why we turned to these guys here.
100
238000
2000
04:15
These are one of the guys I work with. This is a brown capuchin monkey.
101
240000
3000
04:18
These guys are New World primates,
102
243000
2000
04:20
which means they broke off from the human branch
103
245000
2000
04:22
about 35 million years ago.
104
247000
2000
04:24
This means that your great, great, great great, great, great --
105
249000
2000
04:26
with about five million "greats" in there --
106
251000
2000
04:28
grandmother was probably the same great, great, great, great
107
253000
2000
04:30
grandmother with five million "greats" in there
108
255000
2000
04:32
as Holly up here.
109
257000
2000
04:34
You know, so you can take comfort in the fact that this guy up here is a really really distant,
110
259000
3000
04:37
but albeit evolutionary, relative.
111
262000
2000
04:39
The good news about Holly though is that
112
264000
2000
04:41
she doesn't actually have the same kinds of technologies we do.
113
266000
3000
04:44
You know, she's a smart, very cut creature, a primate as well,
114
269000
3000
04:47
but she lacks all the stuff we think might be messing us up.
115
272000
2000
04:49
So she's the perfect test case.
116
274000
2000
04:51
What if we put Holly into the same context as humans?
117
276000
3000
04:54
Does she make the same mistakes as us?
118
279000
2000
04:56
Does she not learn from them? And so on.
119
281000
2000
04:58
And so this is the kind of thing we decided to do.
120
283000
2000
05:00
My students and I got very excited about this a few years ago.
121
285000
2000
05:02
We said, all right, let's, you know, throw so problems at Holly,
122
287000
2000
05:04
see if she messes these things up.
123
289000
2000
05:06
First problem is just, well, where should we start?
124
291000
3000
05:09
Because, you know, it's great for us, but bad for humans.
125
294000
2000
05:11
We make a lot of mistakes in a lot of different contexts.
126
296000
2000
05:13
You know, where are we actually going to start with this?
127
298000
2000
05:15
And because we started this work around the time of the financial collapse,
128
300000
3000
05:18
around the time when foreclosures were hitting the news,
129
303000
2000
05:20
we said, hhmm, maybe we should
130
305000
2000
05:22
actually start in the financial domain.
131
307000
2000
05:24
Maybe we should look at monkey's economic decisions
132
309000
3000
05:27
and try to see if they do the same kinds of dumb things that we do.
133
312000
3000
05:30
Of course, that's when we hit a sort second problem --
134
315000
2000
05:32
a little bit more methodological --
135
317000
2000
05:34
which is that, maybe you guys don't know,
136
319000
2000
05:36
but monkeys don't actually use money. I know, you haven't met them.
137
321000
3000
05:39
But this is why, you know, they're not in the queue behind you
138
324000
2000
05:41
at the grocery store or the ATM -- you know, they don't do this stuff.
139
326000
3000
05:44
So now we faced, you know, a little bit of a problem here.
140
329000
3000
05:47
How are we actually going to ask monkeys about money
141
332000
2000
05:49
if they don't actually use it?
142
334000
2000
05:51
So we said, well, maybe we should just, actually just suck it up
143
336000
2000
05:53
and teach monkeys how to use money.
144
338000
2000
05:55
So that's just what we did.
145
340000
2000
05:57
What you're looking at over here is actually the first unit that I know of
146
342000
3000
06:00
of non-human currency.
147
345000
2000
06:02
We weren't very creative at the time we started these studies,
148
347000
2000
06:04
so we just called it a token.
149
349000
2000
06:06
But this is the unit of currency that we've taught our monkeys at Yale
150
351000
3000
06:09
to actually use with humans,
151
354000
2000
06:11
to actually buy different pieces of food.
152
356000
3000
06:14
It doesn't look like much -- in fact, it isn't like much.
153
359000
2000
06:16
Like most of our money, it's just a piece of metal.
154
361000
2000
06:18
As those of you who've taken currencies home from your trip know,
155
363000
3000
06:21
once you get home, it's actually pretty useless.
156
366000
2000
06:23
It was useless to the monkeys at first
157
368000
2000
06:25
before they realized what they could do with it.
158
370000
2000
06:27
When we first gave it to them in their enclosures,
159
372000
2000
06:29
they actually kind of picked them up, looked at them.
160
374000
2000
06:31
They were these kind of weird things.
161
376000
2000
06:33
But very quickly, the monkeys realized
162
378000
2000
06:35
that they could actually hand these tokens over
163
380000
2000
06:37
to different humans in the lab for some food.
164
382000
3000
06:40
And so you see one of our monkeys, Mayday, up here doing this.
165
385000
2000
06:42
This is A and B are kind of the points where she's sort of a little bit
166
387000
3000
06:45
curious about these things -- doesn't know.
167
390000
2000
06:47
There's this waiting hand from a human experimenter,
168
392000
2000
06:49
and Mayday quickly figures out, apparently the human wants this.
169
394000
3000
06:52
Hands it over, and then gets some food.
170
397000
2000
06:54
It turns out not just Mayday, all of our monkeys get good
171
399000
2000
06:56
at trading tokens with human salesman.
172
401000
2000
06:58
So here's just a quick video of what this looks like.
173
403000
2000
07:00
Here's Mayday. She's going to be trading a token for some food
174
405000
3000
07:03
and waiting happily and getting her food.
175
408000
3000
07:06
Here's Felix, I think. He's our alpha male; he's a kind of big guy.
176
411000
2000
07:08
But he too waits patiently, gets his food and goes on.
177
413000
3000
07:11
So the monkeys get really good at this.
178
416000
2000
07:13
They're surprisingly good at this with very little training.
179
418000
3000
07:16
We just allowed them to pick this up on their own.
180
421000
2000
07:18
The question is: is this anything like human money?
181
423000
2000
07:20
Is this a market at all,
182
425000
2000
07:22
or did we just do a weird psychologist's trick
183
427000
2000
07:24
by getting monkeys to do something,
184
429000
2000
07:26
looking smart, but not really being smart.
185
431000
2000
07:28
And so we said, well, what would the monkeys spontaneously do
186
433000
3000
07:31
if this was really their currency, if they were really using it like money?
187
436000
3000
07:34
Well, you might actually imagine them
188
439000
2000
07:36
to do all the kinds of smart things
189
441000
2000
07:38
that humans do when they start exchanging money with each other.
190
443000
3000
07:41
You might have them start paying attention to price,
191
446000
3000
07:44
paying attention to how much they buy --
192
449000
2000
07:46
sort of keeping track of their monkey token, as it were.
193
451000
3000
07:49
Do the monkeys do anything like this?
194
454000
2000
07:51
And so our monkey marketplace was born.
195
456000
3000
07:54
The way this works is that
196
459000
2000
07:56
our monkeys normally live in a kind of big zoo social enclosure.
197
461000
3000
07:59
When they get a hankering for some treats,
198
464000
2000
08:01
we actually allowed them a way out
199
466000
2000
08:03
into a little smaller enclosure where they could enter the market.
200
468000
2000
08:05
Upon entering the market --
201
470000
2000
08:07
it was actually a much more fun market for the monkeys than most human markets
202
472000
2000
08:09
because, as the monkeys entered the door of the market,
203
474000
3000
08:12
a human would give them a big wallet full of tokens
204
477000
2000
08:14
so they could actually trade the tokens
205
479000
2000
08:16
with one of these two guys here --
206
481000
2000
08:18
two different possible human salesmen
207
483000
2000
08:20
that they could actually buy stuff from.
208
485000
2000
08:22
The salesmen were students from my lab.
209
487000
2000
08:24
They dressed differently; they were different people.
210
489000
2000
08:26
And over time, they did basically the same thing
211
491000
3000
08:29
so the monkeys could learn, you know,
212
494000
2000
08:31
who sold what at what price -- you know, who was reliable, who wasn't, and so on.
213
496000
3000
08:34
And you can see that each of the experimenters
214
499000
2000
08:36
is actually holding up a little, yellow food dish.
215
501000
3000
08:39
and that's what the monkey can for a single token.
216
504000
2000
08:41
So everything costs one token,
217
506000
2000
08:43
but as you can see, sometimes tokens buy more than others,
218
508000
2000
08:45
sometimes more grapes than others.
219
510000
2000
08:47
So I'll show you a quick video of what this marketplace actually looks like.
220
512000
3000
08:50
Here's a monkey-eye-view. Monkeys are shorter, so it's a little short.
221
515000
3000
08:53
But here's Honey.
222
518000
2000
08:55
She's waiting for the market to open a little impatiently.
223
520000
2000
08:57
All of a sudden the market opens. Here's her choice: one grapes or two grapes.
224
522000
3000
09:00
You can see Honey, very good market economist,
225
525000
2000
09:02
goes with the guy who gives more.
226
527000
3000
09:05
She could teach our financial advisers a few things or two.
227
530000
2000
09:07
So not just Honey,
228
532000
2000
09:09
most of the monkeys went with guys who had more.
229
534000
3000
09:12
Most of the monkeys went with guys who had better food.
230
537000
2000
09:14
When we introduced sales, we saw the monkeys paid attention to that.
231
539000
3000
09:17
They really cared about their monkey token dollar.
232
542000
3000
09:20
The more surprising thing was that when we collaborated with economists
233
545000
3000
09:23
to actually look at the monkeys' data using economic tools,
234
548000
3000
09:26
they basically matched, not just qualitatively,
235
551000
3000
09:29
but quantitatively with what we saw
236
554000
2000
09:31
humans doing in a real market.
237
556000
2000
09:33
So much so that, if you saw the monkeys' numbers,
238
558000
2000
09:35
you couldn't tell whether they came from a monkey or a human in the same market.
239
560000
3000
09:38
And what we'd really thought we'd done
240
563000
2000
09:40
is like we'd actually introduced something
241
565000
2000
09:42
that, at least for the monkeys and us,
242
567000
2000
09:44
works like a real financial currency.
243
569000
2000
09:46
Question is: do the monkeys start messing up in the same ways we do?
244
571000
3000
09:49
Well, we already saw anecdotally a couple of signs that they might.
245
574000
3000
09:52
One thing we never saw in the monkey marketplace
246
577000
2000
09:54
was any evidence of saving --
247
579000
2000
09:56
you know, just like our own species.
248
581000
2000
09:58
The monkeys entered the market, spent their entire budget
249
583000
2000
10:00
and then went back to everyone else.
250
585000
2000
10:02
The other thing we also spontaneously saw,
251
587000
2000
10:04
embarrassingly enough,
252
589000
2000
10:06
is spontaneous evidence of larceny.
253
591000
2000
10:08
The monkeys would rip-off the tokens at every available opportunity --
254
593000
3000
10:11
from each other, often from us --
255
596000
2000
10:13
you know, things we didn't necessarily think we were introducing,
256
598000
2000
10:15
but things we spontaneously saw.
257
600000
2000
10:17
So we said, this looks bad.
258
602000
2000
10:19
Can we actually see if the monkeys
259
604000
2000
10:21
are doing exactly the same dumb things as humans do?
260
606000
3000
10:24
One possibility is just kind of let
261
609000
2000
10:26
the monkey financial system play out,
262
611000
2000
10:28
you know, see if they start calling us for bailouts in a few years.
263
613000
2000
10:30
We were a little impatient so we wanted
264
615000
2000
10:32
to sort of speed things up a bit.
265
617000
2000
10:34
So we said, let's actually give the monkeys
266
619000
2000
10:36
the same kinds of problems
267
621000
2000
10:38
that humans tend to get wrong
268
623000
2000
10:40
in certain kinds of economic challenges,
269
625000
2000
10:42
or certain kinds of economic experiments.
270
627000
2000
10:44
And so, since the best way to see how people go wrong
271
629000
3000
10:47
is to actually do it yourself,
272
632000
2000
10:49
I'm going to give you guys a quick experiment
273
634000
2000
10:51
to sort of watch your own financial intuitions in action.
274
636000
2000
10:53
So imagine that right now
275
638000
2000
10:55
I handed each and every one of you
276
640000
2000
10:57
a thousand U.S. dollars -- so 10 crisp hundred dollar bills.
277
642000
3000
11:00
Take these, put it in your wallet
278
645000
2000
11:02
and spend a second thinking about what you're going to do with it.
279
647000
2000
11:04
Because it's yours now; you can buy whatever you want.
280
649000
2000
11:06
Donate it, take it, and so on.
281
651000
2000
11:08
Sounds great, but you get one more choice to earn a little bit more money.
282
653000
3000
11:11
And here's your choice: you can either be risky,
283
656000
3000
11:14
in which case I'm going to flip one of these monkey tokens.
284
659000
2000
11:16
If it comes up heads, you're going to get a thousand dollars more.
285
661000
2000
11:18
If it comes up tails, you get nothing.
286
663000
2000
11:20
So it's a chance to get more, but it's pretty risky.
287
665000
3000
11:23
Your other option is a bit safe. Your just going to get some money for sure.
288
668000
3000
11:26
I'm just going to give you 500 bucks.
289
671000
2000
11:28
You can stick it in your wallet and use it immediately.
290
673000
3000
11:31
So see what your intuition is here.
291
676000
2000
11:33
Most people actually go with the play-it-safe option.
292
678000
3000
11:36
Most people say, why should I be risky when I can get 1,500 dollars for sure?
293
681000
3000
11:39
This seems like a good bet. I'm going to go with that.
294
684000
2000
11:41
You might say, eh, that's not really irrational.
295
686000
2000
11:43
People are a little risk-averse. So what?
296
688000
2000
11:45
Well, the "so what?" comes when start thinking
297
690000
2000
11:47
about the same problem
298
692000
2000
11:49
set up just a little bit differently.
299
694000
2000
11:51
So now imagine that I give each and every one of you
300
696000
2000
11:53
2,000 dollars -- 20 crisp hundred dollar bills.
301
698000
3000
11:56
Now you can buy double to stuff you were going to get before.
302
701000
2000
11:58
Think about how you'd feel sticking it in your wallet.
303
703000
2000
12:00
And now imagine that I have you make another choice
304
705000
2000
12:02
But this time, it's a little bit worse.
305
707000
2000
12:04
Now, you're going to be deciding how you're going to lose money,
306
709000
3000
12:07
but you're going to get the same choice.
307
712000
2000
12:09
You can either take a risky loss --
308
714000
2000
12:11
so I'll flip a coin. If it comes up heads, you're going to actually lose a lot.
309
716000
3000
12:14
If it comes up tails, you lose nothing, you're fine, get to keep the whole thing --
310
719000
3000
12:17
or you could play it safe, which means you have to reach back into your wallet
311
722000
3000
12:20
and give me five of those $100 bills, for certain.
312
725000
3000
12:23
And I'm seeing a lot of furrowed brows out there.
313
728000
3000
12:26
So maybe you're having the same intuitions
314
731000
2000
12:28
as the subjects that were actually tested in this,
315
733000
2000
12:30
which is when presented with these options,
316
735000
2000
12:32
people don't choose to play it safe.
317
737000
2000
12:34
They actually tend to go a little risky.
318
739000
2000
12:36
The reason this is irrational is that we've given people in both situations
319
741000
3000
12:39
the same choice.
320
744000
2000
12:41
It's a 50/50 shot of a thousand or 2,000,
321
746000
3000
12:44
or just 1,500 dollars with certainty.
322
749000
2000
12:46
But people's intuitions about how much risk to take
323
751000
3000
12:49
varies depending on where they started with.
324
754000
2000
12:51
So what's going on?
325
756000
2000
12:53
Well, it turns out that this seems to be the result
326
758000
2000
12:55
of at least two biases that we have at the psychological level.
327
760000
3000
12:58
One is that we have a really hard time thinking in absolute terms.
328
763000
3000
13:01
You really have to do work to figure out,
329
766000
2000
13:03
well, one option's a thousand, 2,000;
330
768000
2000
13:05
one is 1,500.
331
770000
2000
13:07
Instead, we find it very easy to think in very relative terms
332
772000
3000
13:10
as options change from one time to another.
333
775000
3000
13:13
So we think of things as, "Oh, I'm going to get more," or "Oh, I'm going to get less."
334
778000
3000
13:16
This is all well and good, except that
335
781000
2000
13:18
changes in different directions
336
783000
2000
13:20
actually effect whether or not we think
337
785000
2000
13:22
options are good or not.
338
787000
2000
13:24
And this leads to the second bias,
339
789000
2000
13:26
which economists have called loss aversion.
340
791000
2000
13:28
The idea is that we really hate it when things go into the red.
341
793000
3000
13:31
We really hate it when we have to lose out on some money.
342
796000
2000
13:33
And this means that sometimes we'll actually
343
798000
2000
13:35
switch our preferences to avoid this.
344
800000
2000
13:37
What you saw in that last scenario is that
345
802000
2000
13:39
subjects get risky
346
804000
2000
13:41
because they want the small shot that there won't be any loss.
347
806000
3000
13:44
That means when we're in a risk mindset --
348
809000
2000
13:46
excuse me, when we're in a loss mindset,
349
811000
2000
13:48
we actually become more risky,
350
813000
2000
13:50
which can actually be really worrying.
351
815000
2000
13:52
These kinds of things play out in lots of bad ways in humans.
352
817000
3000
13:55
They're why stock investors hold onto losing stocks longer --
353
820000
3000
13:58
because they're evaluating them in relative terms.
354
823000
2000
14:00
They're why people in the housing market refused to sell their house --
355
825000
2000
14:02
because they don't want to sell at a loss.
356
827000
2000
14:04
The question we were interested in
357
829000
2000
14:06
is whether the monkeys show the same biases.
358
831000
2000
14:08
If we set up those same scenarios in our little monkey market,
359
833000
3000
14:11
would they do the same thing as people?
360
836000
2000
14:13
And so this is what we did, we gave the monkeys choices
361
838000
2000
14:15
between guys who were safe -- they did the same thing every time --
362
840000
3000
14:18
or guys who were risky --
363
843000
2000
14:20
they did things differently half the time.
364
845000
2000
14:22
And then we gave them options that were bonuses --
365
847000
2000
14:24
like you guys did in the first scenario --
366
849000
2000
14:26
so they actually have a chance more,
367
851000
2000
14:28
or pieces where they were experiencing losses --
368
853000
3000
14:31
they actually thought they were going to get more than they really got.
369
856000
2000
14:33
And so this is what this looks like.
370
858000
2000
14:35
We introduced the monkeys to two new monkey salesmen.
371
860000
2000
14:37
The guy on the left and right both start with one piece of grape,
372
862000
2000
14:39
so it looks pretty good.
373
864000
2000
14:41
But they're going to give the monkeys bonuses.
374
866000
2000
14:43
The guy on the left is a safe bonus.
375
868000
2000
14:45
All the time, he adds one, to give the monkeys two.
376
870000
3000
14:48
The guy on the right is actually a risky bonus.
377
873000
2000
14:50
Sometimes the monkeys get no bonus -- so this is a bonus of zero.
378
875000
3000
14:53
Sometimes the monkeys get two extra.
379
878000
3000
14:56
For a big bonus, now they get three.
380
881000
2000
14:58
But this is the same choice you guys just faced.
381
883000
2000
15:00
Do the monkeys actually want to play it safe
382
885000
3000
15:03
and then go with the guy who's going to do the same thing on every trial,
383
888000
2000
15:05
or do they want to be risky
384
890000
2000
15:07
and try to get a risky, but big, bonus,
385
892000
2000
15:09
but risk the possibility of getting no bonus.
386
894000
2000
15:11
People here played it safe.
387
896000
2000
15:13
Turns out, the monkeys play it safe too.
388
898000
2000
15:15
Qualitatively and quantitatively,
389
900000
2000
15:17
they choose exactly the same way as people,
390
902000
2000
15:19
when tested in the same thing.
391
904000
2000
15:21
You might say, well, maybe the monkeys just don't like risk.
392
906000
2000
15:23
Maybe we should see how they do with losses.
393
908000
2000
15:25
And so we ran a second version of this.
394
910000
2000
15:27
Now, the monkeys meet two guys
395
912000
2000
15:29
who aren't giving them bonuses;
396
914000
2000
15:31
they're actually giving them less than they expect.
397
916000
2000
15:33
So they look like they're starting out with a big amount.
398
918000
2000
15:35
These are three grapes; the monkey's really psyched for this.
399
920000
2000
15:37
But now they learn these guys are going to give them less than they expect.
400
922000
3000
15:40
They guy on the left is a safe loss.
401
925000
2000
15:42
Every single time, he's going to take one of these away
402
927000
3000
15:45
and give the monkeys just two.
403
930000
2000
15:47
the guy on the right is the risky loss.
404
932000
2000
15:49
Sometimes he gives no loss, so the monkeys are really psyched,
405
934000
3000
15:52
but sometimes he actually gives a big loss,
406
937000
2000
15:54
taking away two to give the monkeys only one.
407
939000
2000
15:56
And so what do the monkeys do?
408
941000
2000
15:58
Again, same choice; they can play it safe
409
943000
2000
16:00
for always getting two grapes every single time,
410
945000
3000
16:03
or they can take a risky bet and choose between one and three.
411
948000
3000
16:06
The remarkable thing to us is that, when you give monkeys this choice,
412
951000
3000
16:09
they do the same irrational thing that people do.
413
954000
2000
16:11
They actually become more risky
414
956000
2000
16:13
depending on how the experimenters started.
415
958000
3000
16:16
This is crazy because it suggests that the monkeys too
416
961000
2000
16:18
are evaluating things in relative terms
417
963000
2000
16:20
and actually treating losses differently than they treat gains.
418
965000
3000
16:23
So what does all of this mean?
419
968000
2000
16:25
Well, what we've shown is that, first of all,
420
970000
2000
16:27
we can actually give the monkeys a financial currency,
421
972000
2000
16:29
and they do very similar things with it.
422
974000
2000
16:31
They do some of the smart things we do,
423
976000
2000
16:33
some of the kind of not so nice things we do,
424
978000
2000
16:35
like steal it and so on.
425
980000
2000
16:37
But they also do some of the irrational things we do.
426
982000
2000
16:39
They systematically get things wrong
427
984000
2000
16:41
and in the same ways that we do.
428
986000
2000
16:43
This is the first take-home message of the Talk,
429
988000
2000
16:45
which is that if you saw the beginning of this and you thought,
430
990000
2000
16:47
oh, I'm totally going to go home and hire a capuchin monkey financial adviser.
431
992000
2000
16:49
They're way cuter than the one at ... you know --
432
994000
2000
16:51
Don't do that; they're probably going to be just as dumb
433
996000
2000
16:53
as the human one you already have.
434
998000
3000
16:56
So, you know, a little bad -- Sorry, sorry, sorry.
435
1001000
2000
16:58
A little bad for monkey investors.
436
1003000
2000
17:00
But of course, you know, the reason you're laughing is bad for humans too.
437
1005000
3000
17:03
Because we've answered the question we started out with.
438
1008000
3000
17:06
We wanted to know where these kinds of errors came from.
439
1011000
2000
17:08
And we started with the hope that maybe we can
440
1013000
2000
17:10
sort of tweak our financial institutions,
441
1015000
2000
17:12
tweak our technologies to make ourselves better.
442
1017000
3000
17:15
But what we've learn is that these biases might be a deeper part of us than that.
443
1020000
3000
17:18
In fact, they might be due to the very nature
444
1023000
2000
17:20
of our evolutionary history.
445
1025000
2000
17:22
You know, maybe it's not just humans
446
1027000
2000
17:24
at the right side of this chain that's duncey.
447
1029000
2000
17:26
Maybe it's sort of duncey all the way back.
448
1031000
2000
17:28
And this, if we believe the capuchin monkey results,
449
1033000
3000
17:31
means that these duncey strategies
450
1036000
2000
17:33
might be 35 million years old.
451
1038000
2000
17:35
That's a long time for a strategy
452
1040000
2000
17:37
to potentially get changed around -- really, really old.
453
1042000
3000
17:40
What do we know about other old strategies like this?
454
1045000
2000
17:42
Well, one thing we know is that they tend to be really hard to overcome.
455
1047000
3000
17:45
You know, think of our evolutionary predilection
456
1050000
2000
17:47
for eating sweet things, fatty things like cheesecake.
457
1052000
3000
17:50
You can't just shut that off.
458
1055000
2000
17:52
You can't just look at the dessert cart as say, "No, no, no. That looks disgusting to me."
459
1057000
3000
17:55
We're just built differently.
460
1060000
2000
17:57
We're going to perceive it as a good thing to go after.
461
1062000
2000
17:59
My guess is that the same thing is going to be true
462
1064000
2000
18:01
when humans are perceiving
463
1066000
2000
18:03
different financial decisions.
464
1068000
2000
18:05
When you're watching your stocks plummet into the red,
465
1070000
2000
18:07
when you're watching your house price go down,
466
1072000
2000
18:09
you're not going to be able to see that
467
1074000
2000
18:11
in anything but old evolutionary terms.
468
1076000
2000
18:13
This means that the biases
469
1078000
2000
18:15
that lead investors to do badly,
470
1080000
2000
18:17
that lead to the foreclosure crisis
471
1082000
2000
18:19
are going to be really hard to overcome.
472
1084000
2000
18:21
So that's the bad news. The question is: is there any good news?
473
1086000
2000
18:23
I'm supposed to be up here telling you the good news.
474
1088000
2000
18:25
Well, the good news, I think,
475
1090000
2000
18:27
is what I started with at the beginning of the Talk,
476
1092000
2000
18:29
which is that humans are not only smart;
477
1094000
2000
18:31
we're really inspirationally smart
478
1096000
2000
18:33
to the rest of the animals in the biological kingdom.
479
1098000
3000
18:36
We're so good at overcoming our biological limitations --
480
1101000
3000
18:39
you know, I flew over here in an airplane.
481
1104000
2000
18:41
I didn't have to try to flap my wings.
482
1106000
2000
18:43
I'm wearing contact lenses now so that I can see all of you.
483
1108000
3000
18:46
I don't have to rely on my own near-sightedness.
484
1111000
3000
18:49
We actually have all of these cases
485
1114000
2000
18:51
where we overcome our biological limitations
486
1116000
3000
18:54
through technology and other means, seemingly pretty easily.
487
1119000
3000
18:57
But we have to recognize that we have those limitations.
488
1122000
3000
19:00
And here's the rub.
489
1125000
2000
19:02
It was Camus who once said that, "Man is the only species
490
1127000
2000
19:04
who refuses to be what he really is."
491
1129000
3000
19:07
But the irony is that
492
1132000
2000
19:09
it might only be in recognizing our limitations
493
1134000
2000
19:11
that we can really actually overcome them.
494
1136000
2000
19:13
The hope is that you all will think about your limitations,
495
1138000
3000
19:16
not necessarily as unovercomable,
496
1141000
3000
19:19
but to recognize them, accept them
497
1144000
2000
19:21
and then use the world of design to actually figure them out.
498
1146000
3000
19:24
That might be the only way that we will really be able
499
1149000
3000
19:27
to achieve our own human potential
500
1152000
2000
19:29
and really be the noble species we hope to all be.
501
1154000
3000
19:32
Thank you.
502
1157000
2000
19:34
(Applause)
503
1159000
5000
ABOUT THE SPEAKER
Laurie Santos - Cognitive psychologistLaurie Santos studies primate psychology and monkeynomics -- testing problems in human psychology on primates, who (not so surprisingly) have many of the same predictable irrationalities we do.
Why you should listen
Laurie Santos runs the Comparative Cognition Laboratory (CapLab) at Yale, where she and collaborators across departments (from psychology to primatology to neurobiology) explore the evolutionary origins of the human mind by studying lemurs, capuchin monkeys and other primates. The twist: Santos looks not only for positive humanlike traits, like tool-using and altruism, but irrational ones, like biased decisionmaking.
In elegant, carefully constructed experiments, Santos and CapLab have studied how primates understand and categorize objects in the physical world -- for instance, that monkeys understand an object is still whole even when part of it is obscured. Going deeper, their experiments also search for clues that primates possess a theory of mind -- an ability to think about what other people think.
Most recently, the lab has been looking at behaviors that were once the province mainly of novelists: jealousy, frustration, judgment of others' intentions, poor economic choices. In one experiment, Santos and her team taught monkeys to use a form of money, tradeable for food. When certain foods became cheaper, monkeys would, like humans, overbuy. As we humans search for clues to our own irrational behaviors, Santos' research suggests that the source of our genius for bad decisions might be our monkey brains.
Laurie Santos | Speaker | TED.com