1
00:00:00,000 --> 00:00:06,390
The primary explanation for the contents of our political beliefs is not our moral values,

2
00:00:06,510 --> 00:00:12,370
is not our support for, say, equality or diversity or authority. It's our political

3
00:00:12,370 --> 00:00:17,270
alliances. It's who we view as our allies and who we view as our enemies.

4
00:00:27,280 --> 00:00:34,320
Hello and welcome everybody to Game Changer, a podcast by TWS Partners. My name is Florian.

5
00:00:34,320 --> 00:00:35,400
This is Simon.

6
00:00:35,620 --> 00:00:36,760
And this is Miriam.

7
00:00:36,980 --> 00:00:42,900
And in this podcast, we explore applications of game theory in business and everyday life.

8
00:00:48,380 --> 00:00:49,300
Hi, Miriam.

9
00:00:49,500 --> 00:00:49,900
Hi, Flo.

10
00:00:50,140 --> 00:00:54,160
I'm really excited about today's episode because we'll be talking to David Pinsof,

11
00:00:54,280 --> 00:00:59,640
and I'm a big fan. He's a behavioural scientist and he's author of the blog with the funny,

12
00:00:59,920 --> 00:01:04,680
interesting title, Everything is Bullshit.  It's a great blog. He explores how people

13
00:01:05,580 --> 00:01:10,840
rationalise and explain their behaviours. And we'll be talking about that also in today's episode.

14
00:01:11,060 --> 00:01:16,140
Sounds very intriguing. And I heard he also has a new podcast called Evolutionary Psychology.

15
00:01:16,600 --> 00:01:20,440
Perfect.  We'll put that link in the show notes as well. And now we'll dive into the episode.

16
00:01:20,700 --> 00:01:21,560
Great.  Let's go.

17
00:01:22,520 --> 00:01:25,300
David, many thanks for joining us on the podcast today.

18
00:01:25,560 --> 00:01:27,380
Yeah, thanks for inviting me.  Glad to be here.

19
00:01:27,780 --> 00:01:33,080
David, we want to talk with you today about how people form beliefs, how people build alliances,

20
00:01:33,080 --> 00:01:40,060
but we thought we'd start with a practical example where we can see maybe how people do that and how

21
00:01:40,060 --> 00:01:46,920
it sometimes has surprising or reveals surprising contradictions. Namely, we'd like to start with a

22
00:01:46,920 --> 00:01:53,200
survey that you've done together with Cards Against Humanity.  Can you explain a bit when you

23
00:01:53,200 --> 00:01:58,200
did that survey? What was the purpose of the survey? And can share a few of the results that

24
00:01:58,200 --> 00:01:58,780
you found out?

25
00:01:58,780 --> 00:02:05,280
Sure. So this survey or this poll was a series of different polls. We did them monthly throughout

26
00:02:05,280 --> 00:02:12,900
2017 to 2018.  And this was one of several sort of surprise gifts or stunts that we gave people as

27
00:02:12,900 --> 00:02:18,860
part of our Cards Against Humanity Saves America campaign. So this was basically in response to

28
00:02:18,860 --> 00:02:24,280
the Trump election. We were all kind of scared about the direction that America was taking.  So

29
00:02:24,280 --> 00:02:29,320
we started the stunt. And so one of probably the most popular stunt that was part of this series

30
00:02:29,320 --> 00:02:35,300
that you might have heard of was when we bought land on the US-Mexico border as a way of protesting

31
00:02:35,300 --> 00:02:40,260
the border wall. And we got some high-powered attorneys to sort of litigate anyone who tried

32
00:02:40,260 --> 00:02:42,080
to build a wall on that land.

33
00:02:42,440 --> 00:02:47,860
Oh, wow. I didn't know about that. So that's, I guess, a subject for a whole different episode.

34
00:02:48,320 --> 00:02:51,920
Yeah. So it was sort of a way of protesting the border wall. And then we did some other

35
00:02:51,920 --> 00:02:56,700
things as well.  So people basically paid for the stunt without knowing what it would get them. And

36
00:02:56,700 --> 00:03:00,000
that was one of the things that got them. And another thing that got them was a year's worth

37
00:03:00,000 --> 00:03:05,640
of public opinion polls.  And this was called Pulse of the Nation. And I directed this poll. So I was

38
00:03:05,640 --> 00:03:08,500
responsible for sort of coming up with the questions and doing some of the preliminary

39
00:03:08,500 --> 00:03:13,380
data analyses.  And my only directive was just to make sure it was funny.

40
00:03:14,840 --> 00:03:15,980
That sounds like a good premise.

41
00:03:16,040 --> 00:03:20,320
Other than that, I could just research whatever topic I was interested in.  Fortunately, at the

42
00:03:20,320 --> 00:03:25,640
time, I was very interested in political alliances and political contradictions and double standards

43
00:03:25,640 --> 00:03:30,300
and inconsistencies in our ideologies and where they come from. So it was a perfect opportunity

44
00:03:30,300 --> 00:03:35,860
for me to really get some good data on these inconsistencies. So we covered a lot of these.

45
00:03:36,240 --> 00:03:41,200
So we would ask participants sort of one question or if they agreed with one statement. And then

46
00:03:41,200 --> 00:03:45,300
there were a few distractor items. And then we would ask them another question that was sort

47
00:03:45,300 --> 00:03:49,720
of inconsistent with the first one to see if we could document the ideological inconsistency.

48
00:03:49,720 --> 00:03:53,620
So just to give you an example. So we first asked the conservatives in our sample,

49
00:03:53,900 --> 00:03:59,320
do you think that nobody deserves a free handout from the government? And then after a few other

50
00:03:59,320 --> 00:04:06,280
items, we then asked them if the government should do more to help small working class towns in

51
00:04:06,280 --> 00:04:11,020
America's heartland. And it turns out that around 60 percent of the conservatives agreed with both

52
00:04:11,020 --> 00:04:15,640
of these statements simultaneously.  Another example would be people should be more respectful of

53
00:04:15,640 --> 00:04:20,660
authority, but businesses should disobey government regulations that they think are unfair.

54
00:04:20,899 --> 00:04:25,460
Another one is we should be more suspicious of foreigners, but we should trust Vladimir Putin

55
00:04:25,460 --> 00:04:30,560
when he said that he didn't interfere in the 2016 election. Someone's on the left.  It's unfair that

56
00:04:30,560 --> 00:04:35,140
corporate CEOs make millions of dollars a year, but it's fair that Hollywood movie stars make

57
00:04:35,140 --> 00:04:39,500
millions of dollars a year. We shouldn't promote negative stereotypes about a group of people based

58
00:04:39,500 --> 00:04:47,880
on where they were born. But people from the South are racist.  Another one would be men and

59
00:04:47,880 --> 00:04:53,740
women are equal in their talents and abilities, but women are better at empathising and multitasking.

60
00:04:54,300 --> 00:04:58,000
So we have a bunch of these and we showed the Venn diagram and we showed that there was like a

61
00:04:58,000 --> 00:05:02,020
striking degree of overlap between people who endorsed both statements with most people being

62
00:05:02,020 --> 00:05:05,500
in the middle of the Venn diagram. And that was a big part of Pulse of the Nation.  We also asked

63
00:05:05,500 --> 00:05:08,780
some other funny questions, but a lot of people remember those double standards and the Venn

64
00:05:08,780 --> 00:05:12,920
diagrams. So of course that when you tell it like that, on the one hand, it's quite funny because

65
00:05:12,920 --> 00:05:18,900
it feels like people are tricked into making statements that are mutually exclusive or that

66
00:05:18,900 --> 00:05:25,260
don't go well with one another. But coming back to our quote research question, what does this tell

67
00:05:25,260 --> 00:05:30,600
us about how people form belief? Are people just inconsistent? Are they not thinking logically?

68
00:05:30,780 --> 00:05:36,100
Or is there anything you can draw from that survey that can shine a light on why people answer that

69
00:05:36,100 --> 00:05:40,080
way? Yeah, so I don't think that people are stupid.  I don't think that people are thinking

70
00:05:40,080 --> 00:05:45,980
illogically or irrationally. I think there is a hidden rationality to it. So I actually cited the

71
00:05:45,980 --> 00:05:49,840
Pulse of the Nation in an academic paper.  It was the first academic paper to cite these results.

72
00:05:49,980 --> 00:05:55,720
And in that academic paper, I argued that the primary explanation for the contents of our

73
00:05:55,720 --> 00:06:02,180
political beliefs is not our moral values, is not our support for say equality or diversity

74
00:06:02,740 --> 00:06:09,100
or authority. It's our political alliances.  It's who we view as our allies and who we view as

75
00:06:09,100 --> 00:06:13,740
our enemies. And so in that paper, I sort of take great pains to try to map the political

76
00:06:13,740 --> 00:06:18,340
alliance structure in the United States, which groups are left coded, which groups are right

77
00:06:18,340 --> 00:06:23,140
coded. And if you reflect on all the examples I showed you, they all involve taking a slogan or

78
00:06:23,140 --> 00:06:27,620
a statement that is designed to advance the interests of one group that let's say is left

79
00:06:27,620 --> 00:06:31,000
coded.  And then you take it out of its proper context and you apply it to a group that's

80
00:06:31,000 --> 00:06:36,520
right coded. And that's where the inconsistency comes from. So for example, corporate CEOs are

81
00:06:36,520 --> 00:06:41,540
right coded.  Hollywood movie stars are left coded. On the right, foreigners are left coded,

82
00:06:41,660 --> 00:06:45,840
but Vladimir Putin, strangely enough, is right coded. And so virtually all of these double

83
00:06:45,840 --> 00:06:50,080
standards came to me while I was trying to just sort of mix and match and pick group targets that

84
00:06:50,080 --> 00:06:54,480
were left coded or right coded and show that the principle is dropped as soon as it's applied

85
00:06:54,480 --> 00:06:58,640
to a right coded or left coded group or the opposite of what was applied previously.

86
00:06:58,640 --> 00:07:03,620
And so I think this is pretty good evidence that our so-called values or principles are actually

87
00:07:03,620 --> 00:07:08,420
tactics or slogans designed to advance the interests of specific groups. And when you

88
00:07:08,420 --> 00:07:12,600
apply them to different groups, it disappears. So I argue in this paper, Strange Bedfellows,

89
00:07:12,660 --> 00:07:17,020
The Alliance Theory of Political Belief Systems, that if you map the political alliance structure

90
00:07:17,020 --> 00:07:21,780
of any society, you can do a really good job of predicting the specific political beliefs that

91
00:07:21,780 --> 00:07:25,900
will arise in that society.  And in addition, you can do a good job of predicting the types of

92
00:07:25,900 --> 00:07:30,340
inconsistencies and double standards you'll see in the political belief systems. And so the

93
00:07:30,340 --> 00:07:36,360
beliefs stem directly from the political alliances and moral values, contrary to a wide variety of

94
00:07:36,360 --> 00:07:40,160
approaches in political psychology, aren't really playing a primary explanatory role,

95
00:07:40,240 --> 00:07:44,800
or at least that's the argument in my paper. And I can understand that what you describe

96
00:07:44,800 --> 00:07:50,520
can lead to these contradictions, but that also would mean you're making a very interesting and

97
00:07:50,520 --> 00:07:54,600
strong statement.  You're not saying normally, I guess people would believe that I form my

98
00:07:54,600 --> 00:07:59,760
beliefs first. And based on these beliefs, my moral beliefs, my moral compass, based on that,

99
00:08:00,140 --> 00:08:05,500
I choose my alliances. But you say it's basically we got it the wrong way around.  We choose our

100
00:08:05,500 --> 00:08:10,720
alliances first. And based on that, we form our beliefs? Yes, that's correct. So one analogy I

101
00:08:10,720 --> 00:08:15,500
make is to the debate over heliocentrism.  So there are two models of the solar system. One

102
00:08:15,500 --> 00:08:20,440
is where the sun is at the centre. One is where the earth is at the centre.  And they can both

103
00:08:20,440 --> 00:08:24,680
predict the motions of the planets. It's just that the geocentric model has to do way more work

104
00:08:24,680 --> 00:08:28,520
predicting the motion of the planets. There are all sorts of loop-de-loops that the planets go

105
00:08:28,520 --> 00:08:33,700
in their orbit around the earth.  These are called epicycles. And so one model is just more elegant,

106
00:08:33,840 --> 00:08:37,960
more parsimonious. It makes fewer assumptions.  And that's just the Copernican model, where it's

107
00:08:37,960 --> 00:08:41,820
just a bunch of concentric circles. And so similarly, I think we have two ways of understanding

108
00:08:41,820 --> 00:08:46,200
the solar system of our political beliefs. One is we put our political alliances at the centre,

109
00:08:46,200 --> 00:08:50,340
and another is we put our moral values at the centre.  And so what I argue in the paper,

110
00:08:50,400 --> 00:08:54,980
this is the argumentative strategy I take, is that if you take this moral values-centric approach,

111
00:08:55,180 --> 00:09:00,560
you wind up with all sorts of weird loop-de-loops and epicycles in your model of political beliefs.

112
00:09:00,760 --> 00:09:04,560
You have all these inconsistencies that are really tough to explain. I presented you some

113
00:09:04,560 --> 00:09:07,980
of them.  There are many more. But if you put alliances at the centre of our political solar

114
00:09:07,980 --> 00:09:11,940
system, all of a sudden, these epicycles go away, and you no longer have to explain them.

115
00:09:11,940 --> 00:09:16,280
They fall out elegantly from the theory.  And so what I argue is that the alliance-centric

116
00:09:16,280 --> 00:09:21,240
view of political beliefs is the more elegant and parsimonious model, and we should favour it.

117
00:09:21,800 --> 00:09:26,120
And do I also understand that then you can also predict or explain

118
00:09:26,120 --> 00:09:30,800
why people draw certain coalitions? How do you get to that point?

119
00:09:31,040 --> 00:09:34,760
Sure. So this part, unfortunately, is understudied at the moment, but we do have some

120
00:09:34,760 --> 00:09:39,200
empirical evidence for at least a sketch of how we choose our allies and how we form alliances.

121
00:09:39,200 --> 00:09:42,720
And so in the paper, I review evidence for three different criteria we use.

122
00:09:42,940 --> 00:09:46,920
One is we choose allies who are like us in some way, who are similar to us,

123
00:09:46,940 --> 00:09:50,600
who might live in the same place, who might worship the same god, who might have the same

124
00:09:50,600 --> 00:09:54,940
skin colour, et cetera. That's one way that we choose our allies.  I mean, there's plenty of

125
00:09:54,940 --> 00:09:59,000
evidence for that, obviously. And then another way we choose our allies is based off of what's

126
00:09:59,000 --> 00:10:04,140
called transitivity. So transitivity, you can summarise with four statements that will probably

127
00:10:04,140 --> 00:10:08,960
make you dizzy, but I'm going to say them anyway.  So the first one is the enemy of my enemy is my

128
00:10:08,960 --> 00:10:14,920
friend. The second one is the friend of my enemy is my enemy. The third is the enemy of my friend

129
00:10:14,920 --> 00:10:19,460
is my enemy.  And the fourth is the friend of my friend is my friend.

130
00:10:19,740 --> 00:10:23,260
Can you just repeat those just one more time? Sure, sure. I'll slow down a little bit this

131
00:10:23,260 --> 00:10:27,660
time.  So the enemy of my enemy is my friend. So this is an ancient proverb. You might be

132
00:10:27,660 --> 00:10:31,520
familiar with it.  If we have a common enemy, that's going to bring us together. That one's

133
00:10:31,520 --> 00:10:35,880
pretty intuitive. The friend of my enemy.  So this is if you're associated with my enemy,

134
00:10:35,880 --> 00:10:40,000
then you're my enemy. So if you're a communist sympathiser, you're a communist. If you're a Nazi

135
00:10:40,000 --> 00:10:45,720
sympathiser, you're a Nazi.  The third one is the enemy of my friend is my enemy. That's if you piss

136
00:10:45,720 --> 00:10:50,960
off my friend, you piss off me. And then the last one is the friend of my friend is my friend.  So

137
00:10:50,960 --> 00:10:55,160
any friend of yours is a friend of mine. If you say, hey, Dave, I really like this person. I'll

138
00:10:55,160 --> 00:10:59,360
probably be inclined to like them because they're a friend of my friend.  Got it. Okay. Makes

139
00:10:59,360 --> 00:11:03,600
absolute sense.  So similarity, transitivity, those are two well-documented ways that we

140
00:11:03,600 --> 00:11:07,980
choose our allies. The third one is a little fuzzier, a little broader and a little more

141
00:11:07,980 --> 00:11:11,680
understudied, but nevertheless, I think it's very strong and very important. This is called

142
00:11:11,680 --> 00:11:16,780
interdependence.  So this is the degree to which we have common interests. For example, if we're both

143
00:11:16,780 --> 00:11:21,600
wealthy business people, we might have a common interest in promoting libertarian ideology that

144
00:11:21,600 --> 00:11:26,360
might lower our taxes and our regulatory burden. So we have common interests.  If we're both Christian

145
00:11:26,360 --> 00:11:30,900
fundamentalists, we might have an interest in increasing the power and status of devout

146
00:11:30,900 --> 00:11:34,860
Christians in society. So we would have common interests there. If we cooperate with one another,

147
00:11:34,960 --> 00:11:38,240
if we trade with one another, countries who trade with each other are less likely to go to war,

148
00:11:38,440 --> 00:11:43,380
then we're more likely to form alliances.  So granted a little fuzzier, but still, I think,

149
00:11:43,760 --> 00:11:48,060
very important and a powerful factor in our alliances. So you could think of this as kind

150
00:11:48,060 --> 00:11:51,780
of collective self-interest, or if we both have the same self-interest, then we're going to form

151
00:11:51,780 --> 00:11:55,740
an alliance. So those are the three criteria that I think there's strong theoretical and empirical

152
00:11:55,740 --> 00:12:00,600
evidence for them.  And fortunately, that's all we need to assume to get off the ground. As soon as

153
00:12:00,600 --> 00:12:05,240
we assume that we have an alliance psychology and we choose allies who are like us, who share our

154
00:12:05,240 --> 00:12:09,680
enemies and our friends, and who have common interests with us, that is all we need to explain

155
00:12:09,680 --> 00:12:15,600
the diversity and dynamism of alliance structures across nations and across time periods. In the

156
00:12:15,600 --> 00:12:20,080
paper, I go through some of that dynamism.  I show how alliances have changed in America throughout

157
00:12:20,080 --> 00:12:24,300
history and how they're different in different countries and in different contexts. And I use

158
00:12:24,300 --> 00:12:28,780
these criteria, similarity, transitivity, interdependence, to explain how these alliance

159
00:12:28,780 --> 00:12:34,740
structures unfold. And the way you describe it with these rules or interactions that you describe,

160
00:12:34,840 --> 00:12:41,160
that only a few, three of these elements you need to explain coalitions or which coalitions form,

161
00:12:41,440 --> 00:12:47,000
to me, this is a very game theory ring to it, because you describe the interactions and then

162
00:12:47,000 --> 00:12:51,900
you can model how the dynamics play out.  And I think you have done the same thing, right? So

163
00:12:51,900 --> 00:12:59,200
you have created game theoretic models to analyse the forming and involving of coalitions. Can you

164
00:12:59,200 --> 00:13:05,100
explain that model a bit? Sure. So I'm currently looking at a particular game called the Simple

165
00:13:05,100 --> 00:13:11,340
Majority Game that I think really captures the logic of what an alliance is.  So why don't we

166
00:13:11,340 --> 00:13:14,580
actually just play the Simple Majority Game right now? It's a three-player game. I think it would be

167
00:13:14,580 --> 00:13:20,980
fun. Here's how it works.  So I have to choose between Flo and Miriam as an ally, right? Flo has

168
00:13:20,980 --> 00:13:24,980
to choose between Miriam and me. Miriam has to choose between me and Flo. Those are the decisions

169
00:13:24,980 --> 00:13:30,360
we all have to make.  And if any two players mutually choose each other, they form an alliance

170
00:13:30,360 --> 00:13:37,820
and win one point. And the excluded third player loses one point. Okay.  And if we each choose one

171
00:13:37,820 --> 00:13:42,140
of the other two players, let's say we form a kind of love triangle and nobody reciprocates

172
00:13:42,140 --> 00:13:46,240
the other's choice, then we all get nothing. We failed to form an alliance. We get zero.

173
00:13:47,060 --> 00:13:51,260
Okay. Are you ready? We have to choose simultaneously and we can only choose one

174
00:13:51,260 --> 00:13:54,580
person, right? That's correct. And so why don't you go ahead and make your choice in your head

175
00:13:54,580 --> 00:13:59,240
right now.  Let's play it right now. Do you all have your choice? Yeah. Okay.  So I chose Flo. Flo,

176
00:13:59,280 --> 00:14:06,380
who did you choose? I chose Miriam. And I chose Flo.  Okay. All right. That is exactly what I

177
00:14:06,380 --> 00:14:11,300
predicted.  You guys have a lot of shared history. You have a lot in common. You know that you're

178
00:14:11,300 --> 00:14:15,320
more familiar with each other and are more valuable partners to each other than you are with

179
00:14:15,320 --> 00:14:21,180
me.  It's okay. I get it. And so for many reasons, you two make very strong allies.  And so it's not

180
00:14:21,180 --> 00:14:25,460
surprising to me at all that you two chose each other. You have similarity. You likely have

181
00:14:25,460 --> 00:14:29,860
transitivity.  You surely have interdependence. It makes sense that you chose each other as allies.

182
00:14:30,040 --> 00:14:33,980
So that's what I think an alliance is.  It's a mutually beneficial coordinated relationship

183
00:14:33,980 --> 00:14:38,900
that comes at a cost to at least some others. And in this particular example, maybe there's

184
00:14:38,900 --> 00:14:43,800
also a bit of a coordination game part of this because we were just thinking, well,

185
00:14:43,800 --> 00:14:47,700
is there a situation, or at least I was thinking, is there a constellation that

186
00:14:47,700 --> 00:14:52,180
shines out with respect to others? And since we are two moderators, that's a natural

187
00:14:52,620 --> 00:14:57,280
connection. And in this coordination environment, that might be a good focal point to go to.

188
00:14:57,560 --> 00:15:01,680
Yes, exactly. So the simple majority game is in fact, in large part, a coordination game. It's

189
00:15:01,680 --> 00:15:05,440
just a coordination game where we have to coordinate on who receives the benefit and

190
00:15:05,440 --> 00:15:09,640
who pays the cost.  And since the person or the group or the subset that pays the cost is smaller

191
00:15:09,640 --> 00:15:14,240
than the subset who receives the benefit, then really it's more about who to gang up against

192
00:15:14,240 --> 00:15:17,540
than who to support. It's more about who we're against than who we're for.

193
00:15:17,740 --> 00:15:20,960
There's a great quote by a game theorist, I forget his name, where he says that

194
00:15:20,960 --> 00:15:25,740
alliances are primarily against and only derivatively for something.  So in larger

195
00:15:25,740 --> 00:15:29,280
groups, let's say we're in a group of 10 and we're ganging up on one or two,

196
00:15:29,480 --> 00:15:34,280
it's much easier mathematically to deconstruct our choices just by saying who we're against

197
00:15:34,280 --> 00:15:39,420
than by saying the eight players or the six or seven players that we're for. It's just simpler.

198
00:15:39,720 --> 00:15:45,480
And so, yes, it's a coordination game on who is going to be excluded or pay the costs relative

199
00:15:45,480 --> 00:15:49,900
to some subset of insiders.  Or you could also think of it as who are the insiders,

200
00:15:50,420 --> 00:15:53,960
who are the outsiders. Nice. So this really can explain a lot,

201
00:15:54,040 --> 00:15:57,120
I guess.  So in the beginning, we were talking about political alliances, but

202
00:15:57,120 --> 00:16:00,620
you also mentioned other alliances, maybe according to religion and so on.

203
00:16:00,620 --> 00:16:08,400
So how does this game theory part or this game theoretic model explain how a very new movement

204
00:16:08,400 --> 00:16:12,440
or a social movement gets traction? Sure, I think it potentially can. So

205
00:16:12,440 --> 00:16:16,700
I'm just thinking about all the different scenarios in real life that might qualify

206
00:16:16,700 --> 00:16:21,480
as a simple majority game.  You could think about forming a political party as a simple majority

207
00:16:21,480 --> 00:16:26,520
game and political scientists have thought about it this way. If we all prefer a policy that benefits

208
00:16:26,520 --> 00:16:30,780
us but inflicts a small cost on the rest of society or on someone else,

209
00:16:31,120 --> 00:16:36,280
then we all benefit by forming an alliance and advancing our political platform. And people who

210
00:16:36,280 --> 00:16:40,080
are excluded from that alliance or who lose the election, they will pay the cost, but they don't

211
00:16:40,080 --> 00:16:43,860
get the benefit.  And so that has the three outcomes that we discussed, forming alliance,

212
00:16:44,140 --> 00:16:48,360
best outcome, failing to form an alliance, second best, and excluded is the worst.

213
00:16:48,740 --> 00:16:53,920
Another really interesting example that has been on my mind a lot, just because it's kind of haunting

214
00:16:53,920 --> 00:16:58,720
to me, is what you might call the witch hunt game. So let's say you two, Flo and Miriam and me,

215
00:16:58,920 --> 00:17:04,460
we're all certain that one of us is a witch and we just have to figure out who it is.

216
00:17:05,040 --> 00:17:09,900
This creates a really interesting strategic scenario because if we're all certain and we're

217
00:17:09,900 --> 00:17:13,859
just going to keep making accusations until we figure out who it is, well, then we each have a

218
00:17:13,859 --> 00:17:19,319
one third chance of being found guilty, right? One of us is guilty. We're going to figure it out and

219
00:17:19,319 --> 00:17:23,099
we start out in the negative. We start out with an expected payoff of negative one third

220
00:17:23,099 --> 00:17:29,220
burned at the stake, right? And the only way to get out of that negative expected payoff,

221
00:17:29,380 --> 00:17:33,840
to unbury yourself, is to accuse someone else of witchcraft, right? And so we're going to assume

222
00:17:33,840 --> 00:17:38,460
that if two players agree that the third is a witch, then that player suffers the cost.  They

223
00:17:38,460 --> 00:17:41,420
get burned at the stake or whatever. And the person doesn't actually have to be a witch,

224
00:17:41,520 --> 00:17:45,300
obviously. The two just have to agree that they are.  So we each start out with negative one third

225
00:17:45,300 --> 00:17:49,980
of the cost, let's say C. So we'll start with negative one third C. If Flo and Miriam form an

226
00:17:49,980 --> 00:17:54,680
alliance against me, then their payoffs will go up to zero and my payoffs will go down to negative

227
00:17:54,680 --> 00:17:58,900
C. And so right there we have the three outcomes that are needed that are ranked in the correct

228
00:17:58,900 --> 00:18:03,240
order. The best outcome is if you form an alliance and you get zero. The intermediate outcome is if

229
00:18:03,240 --> 00:18:07,120
we fail to form an alliance and we have the cloud of suspicion hanging over our heads, in which case

230
00:18:07,120 --> 00:18:11,200
we all have negative one third C. And then the worst outcome is if you're excluded from the

231
00:18:11,200 --> 00:18:16,740
alliance and you pay the cost negative C. So that is a simple majority game.  It has all the ingredients

232
00:18:16,740 --> 00:18:20,700
necessary for being a simple majority game. What's interesting is that the best outcome is zero,

233
00:18:20,860 --> 00:18:24,680
but that doesn't matter. All that matters is that the three outcomes are ranked in the correct order.

234
00:18:24,820 --> 00:18:28,920
And so you have the exact same kinds of alliances and exact same kinds of strategies that will work

235
00:18:28,920 --> 00:18:33,060
in the witch hunt game as will work in any other kind of alliance game, including forming political

236
00:18:33,060 --> 00:18:37,060
parties or figuring out whose stuff we should take and redistribute. So that's what's interesting

237
00:18:37,060 --> 00:18:41,060
about the witch hunt game. And so this suggests one possible avenue by which a social movement

238
00:18:41,060 --> 00:18:46,400
might start is if there's paranoia in society about who is a traitor, who is an enemy, who's

239
00:18:46,400 --> 00:18:51,260
a communist, that can get this witch hunt game started and get people forming alliances and

240
00:18:51,260 --> 00:18:55,840
accusing others and ganging up on others merely to raise their payoffs to zero.  And it's in everyone's

241
00:18:55,840 --> 00:19:00,000
interest to do that, even if they don't really even think that there are witches necessarily.

242
00:19:00,240 --> 00:19:04,460
It's still in their strategic interest to play the game and to accuse others. A similar thing

243
00:19:04,460 --> 00:19:08,180
could occur if something bad happens, a drought, a famine, or flood, and we're looking to point the

244
00:19:08,180 --> 00:19:11,840
finger at someone.  We don't know who that someone is. In that case, people have a strategic interest

245
00:19:11,840 --> 00:19:16,000
in blaming it on their political enemies, and that will often happen in the wake of various crises,

246
00:19:16,000 --> 00:19:21,800
economic downturns, natural disasters, etc. And so you can lean back on your existing strategies

247
00:19:21,800 --> 00:19:25,980
in the event of a witch hunt game.  You're not going to accuse your allies of witchcraft. If

248
00:19:25,980 --> 00:19:29,400
someone's similar to you, you won't accuse them. If someone shares your allies and enemies, you

249
00:19:29,400 --> 00:19:32,640
won't accuse them.  If someone is interdependent with you, you're not going to accuse them. At

250
00:19:32,640 --> 00:19:35,800
the same time, if someone is in conflict with you, you're going to accuse them. There's some

251
00:19:35,800 --> 00:19:39,500
interesting empirical evidence that in real life witch hunt games, so in sub-Saharan Africa where

252
00:19:39,500 --> 00:19:43,580
witchcraft belief is still tragically widespread, there is some evidence that people are more

253
00:19:43,580 --> 00:19:49,000
likely to accuse their competitors, that witchcraft accusations often arise from disputes over wealth

254
00:19:49,000 --> 00:19:52,920
or status, or between co-wives in a polygynous marriage where there's competition over the

255
00:19:52,920 --> 00:19:57,720
husband's resources.  So it looks like in a real life witch hunt game, people are targeting their

256
00:19:57,720 --> 00:20:01,160
competitors or people who are negatively interdependent with them, and that is the

257
00:20:01,160 --> 00:20:04,700
same strategy you would expect in any other kind of alliance game, like an alliance game to form

258
00:20:04,700 --> 00:20:10,220
political parties. Very interesting, yeah. It seems like this simple majority game has a lot of

259
00:20:10,220 --> 00:20:15,580
examples.  Yes, it really does. It's one of the most criminally understudied games, I think,

260
00:20:15,660 --> 00:20:19,840
in the social sciences, and I'm trying to rectify that by writing a paper on it. Fantastic,

261
00:20:20,220 --> 00:20:26,180
and then we're looking forward to more.  Switching the topic slightly, you're also working on

262
00:20:26,180 --> 00:20:32,500
something that might be related, which is humour, and we're very interested in that because it's a

263
00:20:32,500 --> 00:20:38,180
very interesting topic to be studying that, of course. How are you studying it? Are you also

264
00:20:38,180 --> 00:20:42,740
using game theory to do that? The short answer is yes. It's not quite as strategic.  It's more just

265
00:20:42,740 --> 00:20:46,740
about modelling the selection pressures behind it, but yes, I am using math models to try to figure

266
00:20:46,740 --> 00:20:51,340
out what the function of humour is. It sounds like fun. It is a lot of fun.  I remember I originally

267
00:20:51,340 --> 00:20:55,720
had the idea. There wasn't any math. I just had the basic idea, and I mentioned it to someone at a

268
00:20:55,720 --> 00:20:58,780
research conference, and they're like, that's a great idea.  Do you have any math? I'm like, oh,

269
00:20:58,780 --> 00:21:02,800
great. Now I have to figure out how to mathematically model this, or no one will take me seriously,

270
00:21:02,940 --> 00:21:06,760
and so I eventually figured out how to model it, and I actually think it helped to clarify the

271
00:21:06,760 --> 00:21:10,560
ideas in my head and make them more precise and led to some new insights that I hadn't otherwise

272
00:21:10,560 --> 00:21:14,840
had just by trying to model it. This is a slightly different game.  The commonality is it's also a

273
00:21:14,840 --> 00:21:19,840
coordination game, but what's different is that nobody loses in this coordination game. No one

274
00:21:19,840 --> 00:21:23,380
gets burned at the stake, thankfully. It's just we all win together or we all lose together,

275
00:21:23,480 --> 00:21:27,440
and we win by just coordinating by getting on the same page.  The simplest example of this kind

276
00:21:27,440 --> 00:21:31,600
of game that I think humour revolves around is we're both trying to pass each other in a narrow hallway,

277
00:21:31,900 --> 00:21:36,040
and so we have two options. We can either go left or go right. The other person has two options.

278
00:21:36,040 --> 00:21:40,680
They can go left or they can go right, and the two players win if they go in opposite directions

279
00:21:40,680 --> 00:21:45,060
and successfully pass each other and make it to their respective destinations. The players both

280
00:21:45,060 --> 00:21:48,680
lose if they move in the same direction and bump into each other, and they continue losing if they

281
00:21:48,680 --> 00:21:52,920
continue doing that and continue dancing back and forth trying to pass each other in the hallway.

282
00:21:53,120 --> 00:21:56,840
What's notable about that is that when we're in that situation and we're dancing back and

283
00:21:56,840 --> 00:22:01,680
forth trying to pass someone, what arises a lot of the time is laughter.  We find that situation

284
00:22:02,240 --> 00:22:07,360
humorous, and I think that is a very good microcosm of what laughter is and what function

285
00:22:07,360 --> 00:22:12,220
it serves, and I think you can extend the logic of that hallway interaction to most, if not all,

286
00:22:12,260 --> 00:22:17,400
of our jokes and our humorous interactions. All, I think, or most, I think, coordination failures.

287
00:22:17,700 --> 00:22:21,820
Failures of one person or another to recognise what our social conventions are, what roles they

288
00:22:21,820 --> 00:22:26,420
should be playing, who's high status or low status, what words correspond to which things, what the

289
00:22:26,420 --> 00:22:30,280
person meant, what they didn't mean, what the right thing to do is in this situation.  These are all

290
00:22:30,280 --> 00:22:34,720
different kinds of coordination failures, and I think humour revolves around them, and so that's

291
00:22:34,720 --> 00:22:39,220
basically the central idea is that we're constantly playing these coordination games. Our culture is

292
00:22:39,220 --> 00:22:45,160
made up of them, and we need some kind of mechanism for figuring out when a mix-up has occurred, when

293
00:22:45,160 --> 00:22:50,120
we're failing to coordinate, and I think that's what our sense of humour evolved to do. It evolved

294
00:22:50,120 --> 00:22:57,360
to detect these kinds of mix-ups and coordination failures and rectify them, adjust, adapt, signal

295
00:22:57,360 --> 00:23:01,240
them with laughter.  We all know that there's been a mix-up, and we don't have to take this seriously

296
00:23:01,240 --> 00:23:05,000
anymore. We know that we're not dancing with each other or trying to hurt each other when we're

297
00:23:05,000 --> 00:23:08,200
passing each other in the hallway. This is all just a big misunderstanding.  We know that we're

298
00:23:08,200 --> 00:23:11,820
just play fighting. We're not real fighting. We know that this communication mix-up, we shouldn't

299
00:23:11,820 --> 00:23:15,920
treat it as serious communication.  We know that it was just a misunderstanding. That's what laughter

300
00:23:15,920 --> 00:23:20,080
is signalling. It's signalling there's been a mix-up.  I know that you know that I know that there's been

301
00:23:20,080 --> 00:23:25,580
a mix-up, and we can nullify the costs or any costs that might have otherwise occurred from this mix-up

302
00:23:25,580 --> 00:23:30,840
and just treat them as no harm, no foul. And so that's the basic idea of what humour and laughter

303
00:23:30,840 --> 00:23:35,740
are all about. Well, great.  Obviously, this very structured approach to humour

304
00:23:35,740 --> 00:23:42,940
appears to us as Germans. I would like to study it myself. I'm very jealous of your research topic.

305
00:23:43,100 --> 00:23:48,100
Another question, though, you said earlier the commonality is the coordination part, but could

306
00:23:48,100 --> 00:23:55,100
it also extend beyond that? Could humour also be a tool to signal alliance or to test alliance or

307
00:23:55,100 --> 00:23:59,560
be seen in a way in that alliance context? Yeah, I think that's probably true. So there's an

308
00:23:59,560 --> 00:24:03,360
interesting alternative theory of humour that I think is actually quite complementary with my

309
00:24:03,360 --> 00:24:08,120
approach called the encryption theory of humour. And the idea here is that we send encrypted

310
00:24:08,120 --> 00:24:14,140
messages through our jokes that can only be decrypted by members of our subculture.

311
00:24:14,540 --> 00:24:20,400
So getting a joke, that's a kind of decryption, right? So what the function of a joke is,

312
00:24:20,400 --> 00:24:25,340
it is a covert signal of we're in the same subculture, we have the same shared information,

313
00:24:25,740 --> 00:24:31,180
and you don't. So these are inside jokes. And what a joke can do is it can covertly

314
00:24:31,180 --> 00:24:35,360
probe who's a part of your group, who's on your side and who's not without alienating anyone.

315
00:24:35,900 --> 00:24:40,220
So Flo and Miriam can tell a joke that they'll get and I won't get. And that's going to piss

316
00:24:40,220 --> 00:24:45,740
me off less than if they just said, screw you, David, go away. We're better than them.  So it's

317
00:24:45,740 --> 00:24:50,280
a way for Flo and Miriam to covertly establish their alliance without threatening me, or at

318
00:24:50,280 --> 00:24:54,360
least I would be in the dark about what was going on. And so that's the idea there is that some of

319
00:24:54,360 --> 00:24:58,980
our jokes function to covertly probe and test and establish alliances. And certainly a lot of

320
00:24:58,980 --> 00:25:03,960
political humour fits this.  I think another way in which humour intersects with politics is that

321
00:25:03,960 --> 00:25:08,060
the target of a joke, the person who was failing to coordinate, who didn't get it, they're often

322
00:25:08,060 --> 00:25:12,880
seen as a bad social partner. They're not the kind of person that you want to partner up with,

323
00:25:12,940 --> 00:25:17,300
work with, marry, follow in a leadership context, et cetera, because they're just going to have a

324
00:25:17,300 --> 00:25:20,460
bad time coordinating with you. They're going to misunderstand you.  They're going to bump into

325
00:25:20,460 --> 00:25:26,180
things, et cetera. They're a fool. And so often humour has a kind of offensive component where

326
00:25:26,180 --> 00:25:30,740
it's mocking someone and it's belittling their status or their value as a social partner.

327
00:25:30,940 --> 00:25:34,360
And so politics, because it's often about attacking the other side's reputation,

328
00:25:34,540 --> 00:25:39,460
we often try to use political humour and political satire to make the other side look foolish,

329
00:25:39,620 --> 00:25:43,180
to make them look like bad social partners. And that is obviously in our strategic interest,

330
00:25:43,180 --> 00:25:46,240
because if the other side looks like a bad, foolish, or stupid social partner,

331
00:25:46,240 --> 00:25:50,660
well, everyone's going to take your side in the conflict. And so both sides engage in this kind

332
00:25:50,660 --> 00:25:55,480
of political mudslinging and use humour as a political weapon.  So those are two ways in which

333
00:25:55,480 --> 00:26:00,420
humour intersect with alliances. We might use it to covertly establish who is allied with whom,

334
00:26:00,560 --> 00:26:05,200
and we might use it to attack our political enemies. I like the first one much better.

335
00:26:06,460 --> 00:26:12,380
As a final question, maybe, do you have a favourite joke you would like to share?

336
00:26:13,200 --> 00:26:18,980
Honestly, not a huge fan of the joke format. I just find it kind of cliche and tiresome. Although

337
00:26:18,980 --> 00:26:23,740
I was a big fan of Mitch Hedberg.  He's a comedian, sadly passed away, but he had a kind of style

338
00:26:23,740 --> 00:26:28,840
comedy where he would just deliver these one-liners. And the way he delivered them was so

339
00:26:28,840 --> 00:26:33,640
hilarious. It was often more in the delivery than in the content.  And the jokes, they were kind of

340
00:26:33,640 --> 00:26:37,540
hard to describe. They were kind of hard to explain what was funny about them. But I think what was

341
00:26:37,540 --> 00:26:42,000
funny about them is that they were subverting the expectations about what a joke was and what

342
00:26:42,000 --> 00:26:47,760
comedy was.  I'm kind of into those meta-level jokes where I sometimes call them anti-jokes,

343
00:26:47,960 --> 00:26:51,960
where they're sort of jokes making fun of comedy itself. And Mitch Hedberg was an expert

344
00:26:51,960 --> 00:26:56,060
in those kinds of jokes. Let me see if I can remember one.  So one of my favourites was,

345
00:26:56,180 --> 00:27:00,140
he would say in his Southern Drawl, whenever I'm going to shave, I think to myself,

346
00:27:00,320 --> 00:27:06,000
there's probably someone else on the planet who's shaving right now. So I say, I'm going to shave too.

347
00:27:08,660 --> 00:27:12,740
Nice.  I like it.

348
00:27:13,240 --> 00:27:18,080
That one stuck with me for whatever reason. I sort of apply that to everything in my life.

349
00:27:19,300 --> 00:27:23,220
If I'm going out, someone's probably on a podcast right now. I'm going to go on a podcast too.

350
00:27:24,380 --> 00:27:27,900
Yeah.  It feels like you're connected to other people. You don't even know.

351
00:27:28,000 --> 00:27:28,460
Yeah, exactly.

352
00:27:30,020 --> 00:27:35,980
Fantastic. Well, this was great. Thank you very much for your time and explaining all of your

353
00:27:35,980 --> 00:27:38,460
research to us.  It was great having you.

354
00:27:38,540 --> 00:27:40,500
Thank you. It was my pleasure.  Many, many thanks.

355
00:27:42,720 --> 00:27:47,740
We were talking to David Pinsof. He's behavioural scientist, author of the blog

356
00:27:47,740 --> 00:27:53,900
Everything is Bullshit, and co-host of the podcast Evolutionary Psychology, the podcast.

357
00:27:54,700 --> 00:27:59,900
We would love to hear your comments, feedback, or questions. Just send us an email to gamechanger

358
00:27:59,900 --> 00:28:08,400
at tws-partners.com. If you have enjoyed this episode, please send us a review or a rating

359
00:28:08,400 --> 00:28:15,360
using Spotify, Apple Podcasts, or your favourite podcast app. As always, many thanks for listening,

360
00:28:15,600 --> 00:28:19,720
and we hope you'll also tune in for the next episode of Game Changer.