Norms and Development: Interdisciplinary Approach

1
Norms and DevelopmentInterdisciplinary Approach

• Week 2
• Explaining Social Norms I
• Evolutionary Perspective

2
A Story of Mr. RATIONAL

• One day, he had his wallet (50) stolen on the
  street. As he is a very busy and successful
  businessman, its just a waste of time to try to
  get it back. He quickly estimated the cost for
  recovering the wallet is more than 200. He is so
  clever and rational, he didn't chase the thief.

3

• If the thief realizes that hes a rational
  person, he will be a sucker for the thief in
  future

4
A Story of Mr. EMOTIONAL

• He is a successful businessman but known as very
  quick-tempered person. When he had his wallet
  stolen, his temper flared up and started to chase
  the thief. He is not calculating the cost of
  chasing anymore. Hes like an animal just driven
  by anger

5

• If the thief knows that he never gives up
  chasing until he gets a wallet recovered, hell
  never be a sucker for the thief.

6
Passions Within Reason by Robert H. Frank (1989)

• In the world, there are many situations in which
  it is worse to try to be rational.
• Emotion, that are considered to be a
  representative of human irrationality, can
  maximize self-interest that cannot be achieved by
  rational individuals

7

• Huh?
• Being irrational is rational?
• Being rational is irrational?

8
The Limit of Rationality
Action
Consequence at time t

• Total consequences of an action
• the sum of consequence at each time

9
Forward vs. Backward Rationality

• Rationality maximization of self-interest
• Forward Rationality evaluation of the action
  before the action is taken ? with practical and
  logical reasons, the range of maximization is
  myopically limited.
• Backward Rationality evaluation of the action
  long after the action was taken.

10
Action
11
Some Terminological Issues

• The term optimization implicitly assumes
  forward rationality.
• Backward rationality seems to be optimizing. But,
  remember that, when an agent with backward
  rationality takes the action that was optimum in
  the past, it may not be optimum anymore in future
  (e.g., what if the consequence of the action is
  evaluated at the time t 200?).

12
Mechanisms of Backward Rationality

• Genetic Evolution
• Successful genotypes spread in a population via
  Darwinian selection.
• Individual Learning
• Trial and error increases the proportion of
  successful behaviors within individuals.
• Social Learning
• Successful behaviors/traits spread in a
  population via imitation.

13

• Evolutionary processes are the mechanisms
  acquiring the behaviors that are backwardly
  rational and emancipate us from the myopic
  forward rationality.
• This is exactly the reason why we pay attention
  to the evolutionary processes as initial
  candidates explaining human behaviors that
  contradict with the principle of
  forward-rationality.

14

• We do NOT argue that all the human behaviors are
  genetically encoded and selected by genetic
  evolution.
• We do NOT believe in that evolutionary processes
  (in general) can explain everything in the world.
• We are just arguing that, when we find human
  behaviors that look like conflicting with
  self-interest, it is a simple and parsimonious
  research strategy to check the possibility that
  such behaviors are the product of backward
  rationality.

15
Roadmap

• Key issues in evolutionary processes
• Report on Axelrod (1986)
• What lacks in evolutionary perspective?
• Discussion

16
Social Learning Random Imitation

• Lets consider social learning.
• Individuals in the population has a trait (A or
  B). A is more beneficial than B.
• An individual (social learner) randomly picks
  another individual (target) and just imitate the
  targets trait.

17
Random social learning does not increase an
adaptive trait in a population.
18
Social Learning Majority-Biased Imitation

• Lets consider social learning.
• Individuals in the population has a trait (A or
  B). A is more beneficial than B.
• Imagine that an individual (social learner)
  randomly picks three other individuals (targets).
  If majority of the targets has the trait X, the
  learner just imitates it. Otherwise, dont
  imitate.

19
Majority-biased social learning just increases
initially popular trait (but not an adaptive
trait in this case).
20
Social Learning Payoff-Biased Imitation

• Lets consider social learning.
• Individuals in the population has a trait (A or
  B). A is more beneficial than B.
• Imagine that an individual (social learner)
  randomly picks another individual (target) and
  imitate the targets trait only when the benefit
  of the target received in the past is larger than
  the benefit the learners received. Otherwise, he
  doesn't imitate the trait.

21
Payoff-biased social learning increases the
proportion of an adaptive trait.
22

• Social learning is a catch-all word. Be careful
  when you use this word. When we say evolution is
  an analogy of social learning, it means
  payoff-biased social learning.
• Other types of social learning and their roles
  are discussed next week

23
Equilibrium and Path Dependency

• As a result of evolutionary processes, a society
  may reach at stable states in which no one can
  get better outcome by changing the behaviors
  evolutionarily stable equilibrium.
• There may be multiple equilibriums existing. In
  such a case, history matters path dependency.

24
An Example Coordination Game

• Individuals are randomly matched with another
  individual and play the game.
• Evolutionary processes increase the proportion of
  adaptive strategy.

25
Very small initial difference results in totally
different society. Once a society is set on a
different path, it cannot escape from the
self-reinforcing processes
26
Implications of Multiple Equilibriums

• Individuals with totally identical mentality or
  psychology may create totally different
  societies.
• Differences between societies cannot be easily
  reduced into the differences in the psychological
  mechanisms.
• Consider the article in the Guardian. Why the
  people, who are considered to be kind and nice,
  are so messy in queuing?

27
When Socially Harmful Norms Persist or Disappear
An Evolutionary Explanation
Equilibrium B
Equilibrium A
1. Because a society is in stable equilibrium,
small perturbation cannot easily move it to
another equilibrium.
Stable equilibrium
28
When Socially Harmful Norms Persist or Disappear
An Evolutionary Explanation
Equilibrium B
Equilibrium A
1. Because a society is in stable equilibrium,
small perturbation cannot easily move it to
another equilibrium.
Stable equilibrium
29
2. If a norm that was socially beneficial in the
past is transformed into socially harmful norm,
the equilibrium may become less stable. This will
enhance a society moves to another stable
equilibrium.
Equilibrium A
Equilibrium B
30
3. However, a society continues to stay in the
same place until when accumulated perturbation
creates very large shock. It is just because a
harmful norm is still a stable equilibrium.
Equilibrium A
Equilibrium B
31
Summary

• Evolutionary processes have a potential for
  explaining 1) seemingly irrational behaviors and
  2) cultural/societal differences with the single
  principle of backward rationality.
• Thus, evolutionary processes may save us from the
  tyranny of infinite numbers of cultural-specific
  psychological concepts
• We focus on evolutionary processes not because it
  can explain everything but because it is a simple
  and parsimonious starting point.

32
Game Theory vs. Evolutionary Game Theory?

• Q. Forward rationality assumed in game theory is
  like Laplacean-demon. Does game theory make the
  same prediction with evolutionary game theory
  that assumes backward rationality?
• A. Yes (but not always). Just remember that
  evolutionary game theory investigates the dynamic
  processes in which a society changes, and game
  theory investigates where a society reaches as a
  result of dynamic evolutionary processes.

33

• Lets move to Axelrod (1986) that first applied
  evolutionary perspective for explaining WHY
  social norms exist.

34
Whats Wrong About Axelrod (1986)?

• The model makes the critical assumption that a
  players vengefulness against nonpunishment is
  the same as the players vengefulness against an
  original defection. (p. 1102, left column).
• ? If this linkage is removed, social norms cannot
  evolve.

35
FF
PP
Step 1 Step 2 Step 3
cooperate punish punish
cooperate free ride free ride

• PP is better than FF as FF is punished in both
  steps 2 and 3.
• However, PF is better off than PP at step 3. If
  PF dominates PP, then FF can invade and this
  results in the evolution of free riders at the
  step 1.

36
A Basic Model of Social Norm

• Step 1. Norm-obedience
• Step 2. Punishment of norm violators
• Even if a social norm is beneficial to everyone
  (functional), it cannot solve the free-rider
  problem.

37
What Have We Learned?

• Some mechanism that makes punishment (at least)
  less disadvantageous is necessary.

38
Why Free-Rider Problem Was Solved in Kameda et
al. (2003)?

• The game structure is slightly different from a
  standard game of social norms ? the cost of
  punishment is very small in communal sharing
  norm.
• Punishment results in the increase of immediate
  benefit ? non-acquirers can get a portion of
  resource from an acquirer (i.e., looting). This
  does not happen in standard game of social norms.

39

• Another solution is to stop interacting with
  norm-violators (Aoki, 2001).
• If people live in a small group and cannot escape
  to the outside, like small villages in mountains,
  punishment is cheap as it is sufficient to break
  the relations (i.e., helping in danger) with
  norm-violators.
• It doesn't work, however, if mobility gets higher
  or group size gets larger. It is difficult to
  imagine other cases (except for the above two) in
  which punishment is so cheap

40

• Summary and Conclusion

41
What Is Evolutionary Perspective?

• This is a powerful and parsimonious perspective
  that can explain several (seemingly) conflicting
  phenomena with the single principle.
• As three mechanisms composing evolutionary
  processes (individual/social learning and
  evolution) are important forces, evolutionary
  processes can be a kind of first-order
  approximation of social dynamics. See my talk at
  Ringberg.

42
What Lacks in Evolutionary Perspective?

• However, it cannot explain the existence of
  punishment of norm-violators in many situations
  and thus insufficient for explaining social
  norms.
• Furthermore, this approach tends to ignore
  psychological mechanisms underlying social norms.
  For instance, where's a room for (seemingly
  maladaptive) emotion, such as guilt/shame, that
  seem to be crucial for maintaining social norms?

43
Towards Cultural-Evolutionary Perspective

• We tried to explain seemingly irrational
  behaviors only with the single principle. In the
  real world, however, several other forces are
  working.
• Cultural-evolutionary perspective is the
  second-order approximation that provides answers
  to the above questions by incorporating other
  forces. Keep in mind that, however, this
  perspective has the strong basis on
  evolutionary-perspective.