Emotional Expressivity as a Signal of Cooperativeness: An Investigation of Facial Expressions of Emotion in the Ultimatum Game

1
Emotional Expressivity as a Signal of
CooperativenessAn Investigation of Facial
Expressions of Emotion in the Ultimatum
Game
International Conference on Social Dilemmas,
Kyoto, Japan

• Joanna Schug1
• Toshio Yamagishi1
• David Matsumoto2
• Yutaka Horita1
• Kemberlee Bonnett2

• 1 Hokkaido University 2 San Francisco State
  University

2
Emotional Expressions as signals of cooperation

• Cooperation among non-kin may evolve in a
  population when reliable signals of cooperative
  tendency exist, instilling cooperators with the
  ability to selectively seek out other cooperators
  as interaction partners (Dawkins, 1976 Frank,
  1988, 2001Hamilton, 1964)
• Many studies have examined whether cooperators
  can be visually identified
• Physical traits (Takahashi et al., 2006,
  Yamagishi et al., 2003)
• Facial expressions of emotion (Boone Buck 2003
  Frank 1988 Gazzaniga Smylie, 1990 Krumhuber
  et al., 2007 Mehu, Little, Dunbar, 2007 Oda
  et al., 2009 Scharlemann, Eckel, Kacelnik
  Wilson, 2001)
• ? Genuine positive emotion Duchenne smiles

2
3
Duchenne smiles

• Duchenne smiles involve the facial musculature
  (orbicularis occuli) which surround the eyes
• Difficult to intentionally control
• Correlated with subjective positive experience
  (Frank, Ekman, Friesen, 1993 Hess, Banse,
  Kappas, 1995 Keltner Bonanno, 1997 )
• Non-Duchenne smiles only involve the zygomatic
  major muscle (pulls back corners of mouth)
• Possible to intentionally control
• Not correlated with subjective positive
  experience (Bonanno et al., 2002, Keltner, 1995)

?Because genuine expressions of positive emotion
cannot easily be faked, they can reliably signal
cooperativeness
3
4
Duchenne smiles as a signal of cooperativeness

• Altruists/Cooperators display higher levels of
  genuine positive emotion ( Duchenne smiles) than
  egoists
• When talking about personal experiences (Shelley
  Kuhlman, 2007)
• When engaged in natural conversation (Oda et al.,
  2009)
• Reading aloud a fairy-tale (Brown et al., 2003)
• Sharing a reward with a friend (Mehu, Grammer
  Dunbar, 2007)

Duchenne smiles signal cooperativeness
4
5
Duchenne smiles as a signal of cooperativeness
Duchenne smiles as a signal of cooperativeness

• However, Duchenne smiles do not always signal
  cooperation
• Pre-school children playing a Prisoners Dilemma
  game show Duchenne smiles when successfully
  defecting on their partner (Matsumoto et al.,
  1986)
• Olympic athletes show Duchenne smiles when
  defeating an opponent (Matsumoto Willingham,
  2007)
• Duchenne smiles were positively related to
  altruistic tendency only when they were displayed
  in situations requiring sharing, but not during a
  waiting period before the experiment (Mehu,
  Grammer Dunbar, 2007)
• The situation in which the emotion is displayed
  is important

5
6
Can negative emotion signal cooperative intent?

• Can negative emotional expressions signal
  cooperative tendency?
• Moral Disgust One recent study
• found that facial musculature
• relevant to the expression of
• disgust was activated in
• response to unfair offers in
• an ultimatum game
• (Chapman, 2009, Science)
• Many negative expressions of
• emotion can be just as (if not more)
• difficult to intentionally falsify than
  Duchenne smiles (Ekman, 2003 Ekman, Roper,
  Hager, 1980 Porter ten Brinke, 2008)
• ? Same logic holds

6
7
Emotional Expressivity as a signal of Cooperation

• Alternative explanation Emotional expressivity
• Emotional expressivity in itself may serve as a
  marker for cooperative behavior and
  trustworthiness (Boon and Buck, 2003)
• Facilitates mind-reading, allowing for mutual
  selection among cooperators
• Expressivity in itself difficult to falsify
• Emotionally expressive individuals are less able
  to deceive others
• ?Makes defection a bad strategy
• Tendency for cooperators in previous studies to
  show higher levels of positive emotion may be due
  to the tendency for cooperators to openly express
  their emotion
• Previous studies have only examined specific
  emotions in positive/ neutral situations

7
8
The current study

• In this study, we examine the facial expressions
  of emotion of cooperators and non-cooperators as
  they face a negative situation
• Unfair offers in the Ultimatum Game
• We identify cooperators using both a behavioral
  and measurement and SVO
• If cooperators are more emotionally expressive,
  they should display higher levels of not only
  positive but also negative emotion

9
Method

• 20 male participants from Hokkaido University
• Truncated Ultimatum Game (Falk et al., 2003)
• Proposer makes a binary choice
  (500500 or 800200)
• Receiver decides whether to accept or reject the
  offer
• If he accepts the offer, both players receive the
  amount offered by the proposer
• If he rejects the offer, both players receive
  nothing

9
10
Method

• Participants played the Ultimatum Game three
  times (each time with a new player, anonymity
  assured)
• In the first two games, participants played the
  role of the responder, facing offers from two
  proposers
• 1. Intentional unfair offer
• Choices were \500(proposer)/ \500(participant)
• \800(proposer)/
  \200(participant)
• 2. Unintentional unfair offer
• Both choices were \800(proposer)/
  \200(participant)
• ? Intention of the proposer is unclear
• In the last game, the participant played the role
  of the proposer
• Chose between \800(self) / \200(other) and \500
  each

10
11
Identification of Cooperators

• Behavioral identification of cooperators
• Those who selected the fair (500/500) offer as a
  proposer in the third round of the ultimatum game
• ?11/20 participants selected the fair offer
• Social Value Orientation
• Ring Measure of Social Value Orientation (SVO)
  Liebrand, 1984
• Classified as pro-socials and pro-selfs
• ?9/20 participants classified as pro-socials

11
12
Coding of Facial Expressions

• Facial expressions videotaped through the monitor
  display (participants were aware they were being
  recorded, however the camera itself was
  relatively obscured)
• Facial Action Coding System (FACS)
• Only comprehensive, anatomically based system for
  scoring facial expressions
• EMFACS system
• Only codes action units relevant to emotional
  signaling
• Identifies eight emotional categories
• anger, disgust, fear, sadness, contempt,
  duchenne smile, non-duchenne smile, and surprise

12
13
Time spans coded

• We focus the analysis on the time period from
  when the unfair offer was displayed on the
  screen, until the participant decided to accept
  or reject the unfair offer.
• Response phase (approx. 15 seconds)
• Negative event intended to elicit negative
  emotion
• We also coded the waiting period before/after the
  response phase
• Pre- response phase (15-25 seconds)
• Post response phase (30 seconds)
• Because each phase varies in length, we divide
  emotion frequency by the length of each phase and
  use the resulting per-second frequencies in the
  analysis

14
Results (Fair vs. Unfair proposers)
14
Decision-making phase of first two games
(intentional/unintentional) pooled
15
Ultimatum game

• May not be a pure measure of cooperation
• Even those who are solely concerned with their
  own welfare may propose a fair offer if they
  believe their offer will be rejected
• Thus, we qualified the behavioral measure with
  the measure of SVO
• 3/11 fair proposers were identified as pro-selfs
• 1/9 unfair proposers was identified as a
  pro-social
• ? We focus our examination on those who were
  consistently identified as cooperators
  (pro-socials) and non-cooperators (pro-selfs)

15
16
Consistent cooperators
16
Decision-making phase of first two games
(intentional/unintentional) pooled
17
Facial expressions displayed in each phase

• Because we had predicted that differences
  between cooperators and non-cooperators would be
  most pronounced in the response phase (when
  participants faced an unfair offer) we examined
  the differences in the total frequency of
  emotional expressions shown in all three game
  phases
• Pre-response Phase
• Response Phase (Ps faced unfair offers)
• Post-response Phase
• As predicted, we found a significant
  interaction between proposer type
  (cooperator/non-cooperator) and game phase (Pre,
  Response, Post) on the total amount of emotional
  expressions observed.
• F (2,28)11.21, p .0003

18
Facial expressions displayed in each phase
18
Phase
19
Facial expressions displayed in each phase
19
Phase
20
Facial expressions displayed in each phase
20
Phase
21
Facial expressions displayed in each phase
Participants displayed more facial expressions
during the decision-making phase (in which they
faced unfair offers in the Ultimatum Game) than
in the waiting periods before and after the
decision-making phase.
21
Phase
22
Positive (Duchenne Smile)
d.99, p.07,
ns
ns
22
Phase
23
Neutral (Non-Duchenne, Surprise)
d1.28, plt.05
ns
ns
23
Phase
24
Negative (Anger, Contempt, Disgust, Fear, Sadness)
d1.13, plt.05
ns
ns
24
Phase
25
Summary

• As in previous studies, cooperators showed higher
  levels of positive emotion (duchenne smiles)
• However, cooperators also showed higher levels of
  neutral (e.g., non-duchenne smiles, surprise) and
  negative (e.g., sadness, disgust, contempt)
  facial expressions of emotion.
• These differences were most apparent when
  participants were faced with a negative situation
  (unfair offers in the UG)

25
26
Implications

• Previous studies have mainly focused on the
  display of genuine positive emotion as a signal
  of cooperation
• Results of this study suggest that this finding
  may stem from the tendency for cooperators to
  openly display all emotion
• Should lead to higher levels of positive emotion
  in situations requiring cooperation
• Should simultaneously lead to higher levels of
  negative emotion in negative situations

26
27
Limitations

• No incentives to falsify emotion
• Emotions displayed in privacy, no interaction
  with other participants
• Future studies should examine situations in which
  participants have incentives to mask and/or
  falsify their emotion
• Uncertain whether non-cooperators feel less
  emotion, or whether they are masking their
  expressions

27
28
Thank you very much for your attention!