Conformity, Reciprocity and the Sense of
Justice
How Social Contract-based Preferences
and Beliefs Explain Norm Compliance: the
Experimental Evidence
!
"
!
# $
!
&
'
((
(( *
)
+
.
0
)
-
,
"
/
#)
1 " ' & "'
%
♣
!
"
# $
&
'
%
) &
(
'
*
%
+
%
",
& %
&
-../#0
!
!
!
'
'"
#
%
' +
&
0
%
)
!
0
1
2
3
" 45 #
&
'0
(
0
!"
0
0
0
0
%
#$%""&'&#%(& ) * +
♣
1
6
,-
,.
+ 7+8 -..9 ! %
(
:
1
"-../# +
;
%
1
< &
'
a
6:10
(
&
3
0
'
+=3
Department of Economics, University of Trento, via Inama, 5 38100 Trento, and EconomEtica, interuniversity
center of research, University Milano-Bicocca, viale dell'Innovazione, 10 20126 Milano, Italy.
([email protected],
b
Department of Economics, University of Trento via Inama, 5 38100 Trento, Italy ([email protected])
1.
)( !/#(& )%)!0 (&1%(& )"
+
0
%
%
,
0 >
,
"=
0 1
,
,
0 6
,
#
%
(
0
( "3
;
0
0 44?@
0
0 444@
)
(
0 -..A@
0
0 449#
>
0
+
0
;
(
0
0
#
"
%
0
%
#
!
"
#
#
0
1
!
%
0
%
!
*
*
(
B *
*
&
'
0
0
%
!
"
;
0
0
3
-
0
#
!
*
*
1
%
$
;
;
*
"
#
C*!
!
%
*
0
*
(
!
1
!
!
"
#0
%
%
&
:
%
'
0
&
&
'
(
'
*
!
!
%
&
'
,
"-..-0 -../0
,
1
0 -..5#
$
"-..A#
)
%
(0
!
+
0 ( $
) 0
%
0 ( ,
)0
+
!
!
*
%
%
%
0
+
%
"
#
0
"
'# ,
)
&
;
)
D
?
)
%
%
:
0
1
0
D
)
!
6 !
(
8
(
0
%
!
0
%
+
0
0
%
!
&
'
%
%
!
!
(
F
"
;
)
#0
& %
'0
"
%
#
!
(
(
"
#
+
&
'"
#
0
%
!
!
%
)
"
(&
'#0
(
1
0
%
(
(
(
2
3
" 45 #
&
0
+
!
E
'!
"
-../#
2.
0 # 02$&%)# %)!(3 4" )" '5/"(&# 6
:
+
1
"
#
<
0
(
@
0
+
D
)0
!
0
!
>
,
0
+
0
0
0
C
%
"
,
0 49A0
A! 9# <
<
#
0
0
1
;
!
(
0
"
#
B
0
;
3
(
0
$
!
0
:
%
/
#
<
0
(
1
!
+
!
0
;
@
0
;
0
1
0
,
%
0
G
%
" %
%
#
6
"
6
#
) >
+
(
%
"
!
)
# $
;
(
(
0
+
0
0
)
+
%
%
;
+
!
!
0
:
1
0
As in Gauthier’s constrained maximisation theory (1986, 1990,1994) and McClennen’s resolute choice theory
(1990a,b, 1993).
A
1
B
0
;
1
(
0
6>! (
+
0
0
*
+
$
0
D
)
:
(
!
!
0
0
B
0
;
%
;
1
&
'
(
0
%
0
&
0
(
( 0
'
0
%
)
">
0
44-# +
"
$
0
449#
F
(
0
0
(
0
"
&
'
#
0
)
!
+
0
5
0
0
1
!;
!
%
:
2
" 45 #0
!
&
3
'
!
"
#
D
) 1
(
9
0
0
!
-
0
0
(
%
1
3
!
2
03
"
#
)
0
3
0
)
D
)*
!
2
Rather ironically, Rawls’s theory has been vindicated in game theoretical terms by the proof that, given a set
of non cooperative equilibria resulting from the natural evolution of society, the only selection compatible with
both the feasibility condition of equilibria and impartiality (invariance under the personal position symmetric
replacement) is Rawlsian maximin or the egalitarian solution (see Binmore 2005). This means that if one wants
to implement a choice under the veil of ignorance through an equilibrium point that guarantees incentive
compatibility, then one must focus on equality or the maximin solution. This is ironical, because if Rawlsian
principles for institutions were stable in the Nash equilibrium sense – that is, if they provided the only
equilibrium selection mechanism under the veil of ignorance, whereas other principles, like utilitarianism, would
be unfeasible and not implementable – they would be complied with only for self-interested reasons and would
dispense with the emergence of any whatever “sense of justice”. Consider, however, that this argument only
proves that Rawls’s proposal is superior to any other social contract solution under the veil of ignorance if what
is required is making a selection within the set of equilibria emerging because of natural evolution by resorting to
the moral artifice of solution invariance under a symmetric translation of the equilibrium set with respect to the
players’ points of view. Rawls’s general argument, however, could be understood as not imposing ex ante the
constraint that a fair agreement should be confined within the naturally evolved set of equilibrium points.
McClennen (1990) on the contrary has taken this part of Rawls’s theory as the basis for his own approach to the
stability of the constitutional contract. Similarly to Rawls and Gauthier, McClennen thinks that compliance is a
disposition to cooperate conditionally on other players’ cooperation. However, he finds Rawls’s solution weak in
so far as it postulates the sense of justice as an attitude which is engendered by the creation of the well-ordered
society’s institutions, chosen under a veil of ignorance, but it is not a matter of rational choice over dispositions
as such. According to McClennen, Rawls’s approach is exogenous with respect to the mechanism of rational
choice, and he seeks to endogenize the sense of justice through his theory of resolute choice extended to the
context of non cooperative games (McClennen 1993). Resolute choice in these contexts means that a player
undertakes by a decision a disposition that commits himself to forgo, at some later decision node in the game
tree, opportunities for defection which are locally advantageous (so that locally defecting can be dominant). The
reason for doing so is a requirement of consistency with an initial plan, which - when followed by all - permits
players to achieve collectively higher payoffs and to fare better. Of course this idea cannot work if players have
the effective option of defecting at a later decision node where they find that it is locally rational to do so
because of local incentives (for example in a last stage having the Prisoners’ Dilemma structure without
reputation effects). Hence McClennen suggests reform of the theory of rational decisions in games, admitting
that in order to allow this kind of choice over effective dispositions, able effectively to commit players, we
should relinquish hypotheses such as the game tree’s separability into its sub-games, and consequently renounce
the possibility of truncating sub trees and substituting them with their local solutions (when available). He
concludes by giving up backward induction (also when it could provide uniquely determined solutions). Our
opinion is that this reform of rationality criteria is too costly to the theory of game, whereas Rawls’s perspective
is endogenous enough for the endeavor to explain the emergence of a sense of justice as a set of attitudes
governed by a disposition having motivational force on its own (the force of a desire) because it is grounded on
the ex ante decision under a veil of ignorance and it influences the cognitive mechanisms of expectations
formation and motivations formation, leading to a preference capable of commanding a decision behavior.
4
*
+
"
(
%
0
%
(
#0
(
1
7
6>! (
<
)
0
0
0
)
)
1
0
D
:
)
!
0
!
"
#0
0
C
0
0
7
(
B
0
)
B
1
0
;
!
0
&
!
' 1
0
1
%
;
1
0
0
"
0
# 1
1
.
0
0
0
(
1
(
(
(
0
0
%
0
0
(
%
1
0
0
!H!
(
&,
!
G
!
! %
)
IJ
(
K
0
)
;
'
"3
0 E4 #
:
0
+
!
&
1
&
$
1
1
0
( ' " E4E#
) &
&
( '
;
1
'
( ' F
0
(
F
0
!
$
)
0
0
!
0
0
3
<
#
@
%
(!
G
@
#
"
#
@
#
(
0
(
0
(
0
(
(
@
(
#
0
(
(
0
0
(
#
@
0
( 0
0
%
F
0
0
1
0
0
0
%
+
3
)
%
(
!
0
&
'
1
0
7
;
!
0
(
%
-
3
%
)
#
0
%
*
)
!
"
1
0
#
!
"
%
0
#
#
0
0
0
D
)0
0
!
!
@
(
@
(
0
0
0
0
%
1
)
!
@
(
"
(
0
#
0
%
*
F
0
&
(
?
6>
*
'
%
)
:
(
0
$
)
%
G
0
G
D
( ) +
G
(
0
@
B
(
!
1
8
3
"
3.
#
3
0
(
)
)' 0&"(2 ' )# "
1
",
-..5#
0 -..-0 -../@
%
"
,
0
#
3
:
7
;
0
7
;
%
;
3
?
"
C
%
# $
We don’t say that the theory is entirely Rawlsian, since it assumes that in the ex ante decision a social contract
is subscribed on the Nash bargaining solution of the relevant game, whereas Rawls would have suggested the
maximin solution. What we simply say is that the solution given to the norm compliance problem through
conformist preferences is strictly consistent with Rawls’ idea of the sense of justice. However consider that the
Nash bargaining solution in a symmetric bargaining situation implies the egalitarian solution which is also
consistent with Rawls’ maximin. As Binmore shows, this consistency illustrates an essential feature of the
decision under the veil of ignorance, when it is restricted to the payoff space resulting from the symmetric
translation of the equilibrium set with respect to the Cartesian axes representing the players’ payoffs (Binmore
2005).
E
0
!
"D
)#
0
0
"
#
"
;
0
0
# +
!
7
&
0
('
0
%
1
%
D
) !
G
$
1
%
!
1
;
#
&
'
!
"
!
0
1
%
0
0 %
%
(
1
!
1
)
:
;
0
+
0
( 0
G
0
G
1
0=
$
<
0
"
#
/
1
D
)
+
0
( 0
0
G
%
G
0
G
0
%
%
:
!
,
-../#<
0
!
"
10
6
1"
#
!
0
σ0
1
σ
1
1
7
'
"
1
1
#0
1
"7$6#
"
&
7
E
0
%
0
) %
0
G
1 1
%
0
"G
%
%
#
."
0
1
#
1
4
+
"
G
0
)
%
)
%
%
#
The Nash Bargaining Solution may be understood as a formal model for the “social contract” that players
would agree in an ex ante (possibly hypothetical) collective decision on the rules that should constrain (at least
as a matter of “ought”) the allocation of surpluses arising form their interactions (see Sacconi 2000, and Binmore
1998, 2005) .
A
0
%
%
10
)
G
1
"
%
" %
0
#
G
%
"
0
G
"
#
1
."
%
#
# %
%
0
1
G
%
%
%
1
#
%
0
!
λ
%
)
/
%
%
1
.
"
#
0
%
0
"
λ
%
0
%
#
λ
)
%
%
10
%
%
G
10
) " %
λ
0
G
#
0
G
1
0
)
0
"
:
;
#
σ"
0
#L
"
%
σ
)
σ! #0
#
Vi (σ ) = Ui (σ ) + λi F [T (σ )]
5
This assumption corresponds to Rawls’s assumption of a capacity to form a sense of justice derivable from the
lower-level moral sentiments.
5
$
)
λ
%
σ@
%
@
"
σ@
7$ #0
%%
!
)
σ0
)
!
) %
"
σ
4.
' )# "&(/%(& ) (3
%& '
B
)
#
7#$/"& )8%0
(
%
0
F
8%
,
1
1
(
(
0
0
(
G
(
%
%
0
(
C
(
+
8%
#
"
"
"
# +
0
?#0
0 d ih =
30
0
"
0
(
C@
0
-#
3
<
,
(
30
9
R
2
0
!
d il =
0
i = {1,2} 1
R
0
3
)
3
%
0
0
(
)
:
8%
,
6
d 2l
d 2h
d 1l
R R R
, ,
3 3 3
R R R
, ,
3 2 6
d 1h
R R R
, ,
2 3 6
R R
, ,0
2 2
#0
)
3"
@
(
30
(
3
30
-
,
(
! "3G?
3G?
?
0
%
0
;
(
,
!
0
R
2
91
0
0
(
0
1
0
( ;
dh =
%
%
B
)
!
#
7
)
(
;
,
%
0
%
+
3
1
8%
&
0
s = R − d 1l + d 2l = R
3GA :
)*
0
<
%
0
%
4
" (
%
0
%,
%
# :
C
'
+
8%
:
,
0
!
!
0
7
"
7
#<
T (σ ) = ∏ (U i (σ ) − ci )
"A#
i∈I
00
(
B
0
;
7
0
"5#
& ,
T
"9#
MAX
(
(
l
l
d1 , d 2
h
T d ,d
T
MIN
-
(d
l
&
)
,
.
& ,
h
,
3
) = T (d , d )
,d
&
R R R
R
=
⋅ ⋅
=
3 3 3
27
l
h
-
h
) = N (d
3
R R R R
= ⋅ ⋅ =
3 2 6
36
h
h
1 , d2
) = R3 ⋅ R3 ⋅ 0 = 0
F
%
"
;
/
-.
0
-# +
'
%
",
(
0
494A#0
%
7
;
%
)
:
d 2l
d 2h
d 1l
R
R
+ λ1 , + λ 2
3
3
R R
,
3 2
d 1h
R R
,
2 3
R R
,
2 2
!
, 1
/
2
'
0
(
)
0
;
7
;
0 " d1l , d 2l #
+
λ
;
;
3
(
)
(
)
R R R
− =
2 3 6
(
)
(
)
R R R
− =
2 3 6
"4#
V1 d 1l , b11 = d 2l , b12 = d 1l > V1 d 1h , b11 = d 2l , b12 = d1h ⇔ λ1 >
" .#
V2 d 2l , b21 = d 1l , b22 = d 1l > V1 d 2h , b21 = d 1l , b22 = d 1h ⇔ λ 2 >
6
" d1l , d 2l #0
<
)
d il !
dl
!
d ih 0
+
%
λ
=
6
%
See also Sacconi and Grimalda (2005b) for details.
-
;
-
(
;
3
1
8%
0
;
,
+
0
0
7
0
0
%
0
;
" d1h , d 2h #
;
<
%
-
0
%
+
!
0
"
-
!
"
#0
!
G
#0
!
+
0
"
%
"
-#0
?#
%
5.
1
72 &0 )(%$! "&:)%)!2 # !/
%
(
=
F
'
8
!
4''5
0
/.
M/
0
,
<
0
%
:
0
+
%
8%
F
,-
5
/
!
+
,
'
1
5
8%
4
,? ,
,
1
0
,-
7
Participants were all students at the University of Trento (mainly from economics, law and sociology courses),
recruited by responding to ads posted in the various departments.
--
"3N M -#
0
+
(
0
" 0 -#0
:
< -/O0 ??O
)
"
-0
/.O
)
3
3!" P -#
?#
1
0
:
0
1
,?
1
0
%
1
,
!
%
1
)
G
, 0 ,-
)
9
<
(
G
6
'
,-
(
/
8;
8A
+
- <;=>?
.<-->?
, <=@>?
-<;=>?
?0?0 A
?0 E0 /
?0 A0 ?
.<-->?
E0 ?0 /
E0 E0 E
E0 A0 -
,<=@>
A0 ?0 ?
A0 E0 -
A0 A0 .
-
F
0
0
0
!
1
0
7
0
%
0
+
0
(
This is an application of the procedure known as the random lottery incentive system (Starmer and Sugden
(1991) and Cubitt et al. (1998)). On adopting this procedure, the round in which the subject has played
occupying the G1 or G2 role is selected with a probability of 2/3, which is the same probability of being
extracted as G1 or G2 in a one-shot version of the game. Note that, if we look at the third phase of the
experiment, using this mechanism we can always compare the choice of each of the players who in that phase
have an active role with his/her choice in the first phase.
9
-?
0
;
"
E#
0
:
(
:
0
%
0
:
4
Figure 4: Second phase. Principles and rules
PRINCIPLE 2.
Principles
PRINCIPLE 1.
“Every player should share the
“Players
benefits; in particular, the player
decisional role could claim a higher
who has not been able to choose
share of benefits”.
Rules
under
a
8A 8; 8-
# ??O -/O E-O
- # /.O -/O 5O
- # -/O ??O E-O
- - # ??O /.O 5O
? # ??O ??O ??O
- ? # /.O /.O .
?0
0 , 0 ,-
1
,?
0
+
0
@
9
play
should not receive less than others”.
8A 8; 8-
+
who
See Appendix 2 for a detailed description of the voting procedure.
-E
G
1
0
%
0
, 0 G
E
0
0
+
0
DE)
(
%
%
.
)
:
%
!
/
73 '
PHASE 1
SUBJECTS ARE DIVIDED INTO 5 GROUPS
OF 3 MEMBERS EACH.
6.
ROUND 1
ROUND 2
ROUND 3
Within each group:
Within each group:
Within each group:
- G1, G2 and G3 roles are assigned.
- G1, G2 and G3 roles are assigned.
- G1, G2 and G3 roles are assigned.
- G1 and G2 play the game
- G1 and G2 play the game
02&
&#%$2 !&#(& )"
- G1 and G2 play the game
PHASE 2
VOTING ON
SUBJECTS ARE DIVIDED
VOTING ON SPECIFIC RULE
GENERAL PRINCIPLE
INTO 5 GROUPS OF 3
( MAX. 10 TRIALS)
( MAX. 5 TRIALS)
MEMBERS EACH.
PHASE 3 (WITHIN EACH GROUP)
G1, G2 AND G3
ROLES ARE
ASSIGNED
G1 AND G2 PLAY THE
PREDICTION
GAME FOR THE FIRST
BY G1 AND G2
TIME
(IMPLEMENT THE RULE
OR ALTERNATIVE
F
%
%
C
1
;
%
E
+ λ1 Q9G?
DE)
10
DA)0
We asked the player to indicate the cell of the payoff matrix in which s/he thought the game would end. In this
way we avoided explicitly asking for his/her opinion about the opponent’s willingness to conform with the rule.
Fe begin by applying the Nash bargaining function to the outcomes of the game used in experiment (2). The
‘fairness values’ given by the function corresponding to the various states resulting form playing each strategy
combination of the game are :
T(4,4)=TMAX = 64
T(3,4)=T(4,3)=60
T(3,3)= T(3,6)=T(6,3)=54
T(4,6)=T(6,4)=48
-/
- 1
λ i Q9G?
"E0 E#
." 0-#
;
)
$
0 G
0
DA) $
"A0 A#
-0
;
:
E0
)
+
0
0
! %
1
0
"
%
# 8
λi 0
F
%
0
0
%
!
!&#(& )A98
0
;
(
MA "/.O
#
9
:
;
<#
+
8%
1
(
0
8%
(
,
0
1
0
T(6,6)= TMIN = 0
0
%
0
0
)
%
1
%<
-
.
,
-
?0?
?P"?GE#λ 0 EP"?GE# λ-
?0A
.
EP"?GE# λ-0 ?P"?GE#λ
EPλ 0 EPλ-
E0A
,
A0?
A0E
A0A
-A
,
:
0
<
0
0
1
0
0 ;
!
!
7
,
&
0
'
8%
7
)
"
D ;
D
)
6
0
#
+
0 D ;
D
)
0
) ,
0
!
0
0
@
0
;
F
F
;
(
0
D
)
0
0
0
;
;
0
0
"F
&
'
0
0
0
# 1
0
0
;
7
%
+
!
&
(
'
0
-5
0
!
!
0
)
%
!
0
7
B
0
;
0
(C
F
D
)
!
!
"
#0
-
1
0
%
!&#(& ) ;98
9
:
+
0
!
"
%
#0
%
0
+
%
0
;
0
DE) F
%
0
0
%
0
0
;
0
12
Note that this implies that only a subset of the observations consistent with the theory may have crucial
discriminating force amongst different theoretical hypotheses on rational action, and we are mainly interested in
producing exactly this kind of evidence. This would also have justified us in placing somewhat more stress on
the ethical nature of the second-phase decision in order to test the level of conformism in the third phase.
-9
!&#(& )-9/
#
!&#(& ) .9(
;
%# 2
,
9
69
9
69
,
;
%#
,9 "/$("
<&
4
&
6
F
)
? +
0
"
,-
1
#0
"
0
<
@ "
#
G
# G
G
G
,
,
?@
G
,-
@
%
)
/.
"/4 ?O#
(
"1
MA
,
#
&3 4
& .&7=
R" M#
?!?
?!E
E!?
E!E
E!A
A!E
A!A
?!A
A!?
B
.
E
/
-.
-.
94
.
/.
R1
O
. ..O
. A5O
- A5O
. ..O
? ??O
? ??O
/4 ??O
. A5O
. ..O
.. ..O
,
-4
,-
,-
0
,
,-0
(
MA
, ",-#
2
ME
,- ", #
0
:
;
0
0
0
0
:
..
,
,-
(
<
..
(
ME0
/A
(
MA
EE
-
,3 1
#
&
63
.&==
23%" -
23%" A B
(
?!E
?!A
E!?
E!E
E!A
A!E
A!A
ME
(
MA
.
.
.
E
/
-..
A
A
;
A;
AA
A;
,A
A@@
5
5
?4
=,
R1
,
,-
1
"
#
?"
#
0
& (
;
ME
"8!
0
(
#
& (
M A'
J!
"
7
)
!;
N--0E
(
!
;
"1
0
N0 !
!
" !
' F
8!
(
(
#
!
MA
(
(
(
(
ME
(
M AK
?#0
M EK
N- - E !.A#
?.
MA
J8!
0
?
? B
1
? 8!
!
? "
#
3%" A
3%" "C ' D.
"C ' D,
#3 &#
-
?
A=
#3 &#
?-
/?
E=
..
=,
A@@
F
%
)
?C
<,
4
+
-0
0
0
/.
•
?-
"4A
#
•
/
" E/
#
•
-
"A
•
;
"?
<
"??O0 ??O0 ??O#0
-
#
"/.O0 /.O0 .O#
-
#
" /.O0 ??O0 5O#
-
" ??O0 /.O0 5O#
=
%
"
#0
1
%
;
"
#0
F
0
•
E?
•
3
AE
"??O0 ??O0 ??O#
"/.O0 /.O0 .O#
-4
"
+
?
1
?.
E#
?0
B
0
5EO
?
)
%1
)
%?
0.&==
&"((&0
8%
"
3
-- -- --
#
#
+
E?
E-
>
AE
A
-4
-9
1B1
=@ =@ @
"
+
>
1B1
-- =@ A*
=@ -- A*
?.
+
>
.
.
1B1
-
1B1
E
+
>
<6
>
-
#
1
)
(
(
ME
)
0
<
Pr(Choice3 = 1) = Φ(α 1 Rule + α 2 Choice1 + α 3 Belief )
4
;
6
;
?
?
& (
& (
M E'@ 4
%
&& (
0
G
(
MA
?-
M E'
M A'@ )
MA'
&
& (
?
(
ME
$
)
(
?
ME +
0
D (
M E)
%
(
? 1
(
"1
7
6 /
ME
/#
'
5E
3
". 5/5#RR
!. -5
". -4/#
$
- 5/
". /?E#RRR
! 55
" 9 #
6
3-
. A4.
6
Q S-
. ....
7
..
>
C< 4
(
0 (
6
ME
?
4
(
(
D (
?
G
RR
7.
0
(
(
ME
&3
0
) 3
ME)
.
?
/O@ RRR
O@
(
&"#/""& )%)!# )#$/"& )"
F
(
A
0
!
+
#
1
#
1
0
<
DA)
0
0
??O0 ??O#
"??O0
-
??
#
0
?
%
&
(
0
1
'
0
0
G
#
1
)
)
G
@
0
"??O0 ??O0 ??O#
DE) "
#
?#
:
DA)
;
0
@
0
0
:
1
E
)
0
(
0
0
7
0
0
%
"
#
"
# 1
0
%
F
0
0
<
#
;
!
"
0 444#
%
F
;
13
C1
This hypothesis is supported by the replies to the debriefing questionnaire.
?E
%
&
;
!
'
0
(
0
0
;
!
%
(
(
0
"
0
#
"
)
0
;
# D>
)
!
#
0
D
(
) "3
0 44?0
(
0 -.. #
0
(
0
0
7
(
"
#0
(
7
*
0
%
)
*
1
0
T
0
%
3
0
%
D
)0
)
0
(
0
6
D
3
#
0
%
3
)
0
)
%
1
1
)
3
%
#0
"
%
!
%
!
( 1
=
%
0
!
(
*
?/
0
%
!
"
>
λ#
%
0
%
0
%
0
!
F
0
0
%
0
:
(
G
1
G
B
0
0
%
:
%
0
3
!
0
0
F
(
%
!
3
)
"
!
%
-#
G
"??O0??O0??O# :
0(
0
0
3
+
(
F
0
G
(
)
C:
7
G
0
"??O0??O0??O#
1
%
1
0
;
0
0
%
+
?A
(
0
%
%
B
0
(
44E0
"3
49.0 $
-..9# 1
%
0
G
;
(
:
!
:
0
G
%
$
:
D
)
)"
$
-..A#
+
0
)
G
$
0
)
0
0
1
G
)
)
0
%
1
%
!
!
"
#
1
%
G
2
0
%
:
)
0
%
%
0
0
D
)!
D
?5
(
0
"
#
)
*
1
&
'
"??O0??O0??O#<
0
(
+
0
*
F
3
0
0
"??O0??O0??O# !
%
?9
38 8387 8
$
0
" 44E#0 &1
$
0
"-..A#0 ?
=
$
8
0
6 0
"-..A#0
9 ?505!E9
@
06
3
(
0 V " 449#0 /
$
0 V "-../#0 0
$
0 , 8 " 44 # :
'
)
$
0 , 8
1
03
>
+1 6
<6
9A
0V
0 > " 44E#0 U
0 W
B%
F
0
=
'
0 3
'
+
03
0
< 1
$
06 "
#0 (
9 0 6
'
0+
)
9)
!
8
0
3
A
"
0 20 6
0 >
3
'0 (
0,
'0 1
3
0
0 B%
1
0 > " 44.#0 U8
? CD0 2 7
(
8
W
" 444#0 &: 1
'
E0 9 5!9A9
B
,
8;
)
0 = "-.. # 1
F ( 6
7
,
< 1
"-..5#0 &7
'0 +
0:
?
(
0 '
0 > " 49A#0 2
=
= 6 0 B%
0 3 " 449# &B
'
< /< ?
,
0
0
0$
0 : "-...# &:
)
..0 AA! 4?
0
1
08
"
$
0 6 " 44-# 1
8
@
9 0 .4A! ?A
0
'0 '
(0 :
0
0 B%
B (
'
8
?
<
"
0
0
,
U0
6
$
,
,
6
#'
< : 8
9)
3
6
+
!
0 B%
3
9'
3
U0 !
/
U
=6
'
,
"-..-#0
9
?4
6
0 B%
?
0
0 8 " 494# &6
?
0 0 A.!54
! +=
'0
//
;
9
,
0 ,
3
0
"-../# U1
'4
0 8 " 44.# &
8
'0 4
3
0
'
/
8%
0
0 8 " 44.#0 )
@
0 8 " 44?#0 &3
0
E07
0 2 " 45 #0 1
3
3 " 49.#0 &:
0
8
1
0
,
3
8;
$
0
6
'04
06
6
"-..9#0 &:
,
9 ! GG
H
8
' 1
)
-G./ >
7
9 ?0 9 ! ?-
8
E'
(
"
<:
#
:
C
'
>
3
"
0 W
A0 7
0 3 " 44 #0 &>
C: 8%
+
3
E.
+
7 ! !6
08
8
7
=
?
F3 1 '
=
8
0
0
0
3
1
'01
"-../#0 4
0>
6
$
6
= 6< B%
, "-..5#0 &+
'0
1
<
-0 4E! 9
>
,
B
A"?#0 -E4!-5A
=60
8
0 B%
0
!6
7
0 A5!44
07
0
4
" 44?#0 &+
)
9?"/#0 -9 ! ?.-
3
7 !
'
' 8
!
' 1
0 /4!9
+
'
)
8
1
0 9 0 45 !459
<
22 )!&7A (3 # )' 0&"(2 ' )# 0 ! $
1
%
?
G = { I , S ,U }
E
0
)
Σi := Σ ( Si )
0
i∈I
$
Σi := × Σi
)
σ ∈Σ
0
U i (σ )
$
0 S = × Si
8
#L &
;
# σ0
"
i∈I
'
)
!
"
#
/
<
Vi (σ ) = U i (σ ) + λi F T (σ )
" #
λ
1
λ >.
%
%
"
D
%
#
#L
)0
T : U i (σ ) →
0 ,
!
%
0
%
)
%
",
+
1
1
0
6
(
,
494#
9
"
9 B := ∆ ( Σ −i ) 0
>)
1
i
#
)
+
σ
bi1 ∈ Bi1
0
B−1 i := × ( B j )
0
0
j ≠i
)
)
( )
B := ∆ B− i 0
2
i
0
<
b ∈B
1
2
i
>
2
i
bi1
I
+
!
0
%
<
/
1
#
3
14
15
"
;
I
0
G
%
>)
.
* #0
+
Grimalda and Sacconi (2002) elaborate on Rabin (1993) in order to define the model of reciprocal conformity.
See Grimalda and Sacconi (2002, 2005) and Sacconi and Grimalda (2007) for details.
E
0
I
>#
(
f i σ i , bi1
"-#
b i1
)
1
(
1
0 T σ i , bi
0T
)
)
) = T ( (b )) − T ((b ))
T σ i , bi1 − T MAX bi1
MAX
MIN
0 T
>)
(b )
MIN
1
i
1
i
(b )
MAX
1
i
%
1
i
1
σ0
1
G
)
)
'
1
#
3
>D
"
0
#
G
"
I
(
b i1
(
)
)
( )
( )
~ 1 2
T bi1 , bi2 − T MAX bi2
f j bi , bi = MAX
T
bi2 − T MIN bi2
"?#
I
( )
"
>)
2
i
#0 b
"
%
1
L
>)
>#
λi 1 + f
"E#
j
(b
2
i
, b i1
)
1 + f i (σ i , b i1
)
<
G
%
! #0
!
;
>#
G
.
%
(
0
0
%
>
0
λ $
0
0
0
%
λ"
(
&−
#
1
%
"
&− "
A0
− # 1
;
0
λ
#
(
0
)
;
0
"/#
1
Vi (σ i , bi1 , bi2 ) = U i (σ i , bi1 ) + λ i 1 + f j ( bi2 , bi1 )
λ #0
E-
"
"
#
1 + f i (σ i , bi1 )
λ
σ
0
)
%
$ σ9
;
0
22 )!&7;9 (&):
# !/
"
/# 8
G
G
0
1
%
(
!
1
& 3!
(
& +> '
&,
%
(
%
(
E?
)
'
&6
0
%
)
'
1
)
+
0
A
%
(
1
%
0
1
:
?
8
%
0
(
-
1
1
%
!
0
, 3 !
!
6 3 !
&
:
16
0
(
,
0
%
C
This made it impossible to identify the members of a particular group by exploiting the information about the
outcome of the voting procedure.
EE
22 )!&7-9 3 ! )%0&#" '"/F5#("G
#3 &#
84 choose to ask for ‘6’ at least once in the first phase
Phase I
Phase II
Phase III (1st time)
Phase III (2nd time)
50 choose rule 33%-33%-
28 choose rule 50%-50%-
33%
0
30 implement
20 do not implement
27 implement
1 do not implement
28 expect ed the
12 expect ed the
25 expect ed the
0 expect ed different
same behavior from
same behavior from
same behavior from
behavior from the
the opponent
the opponent
the opponent
opponent
28 implement ed
22 do not implement
27 implement
27 expect the same
15 expect the same
25 expect the same
0 expect different
behavior from the
behavior from the
behavior from the
behavior from the
opponent
opponent
opponent
opponent
E/
1 do not implement
EA
!
"#
$
%
&
!
! "
"
!
!
'
(
)
*
+
,
*
-
.
!
#
/
!
.
! 0
- )
$
!
1
/
#
1
%
&+
0
'#
!
#
*%
1
.
)
" (
&
$ 2
%
.
1
! "
/
/
!
'
$
! "
)
%
#
*
/
'
(
%
!
!
)
)
$
3
4
& 1 6
789:
$
5
4
.
(
%
5
!
(
%
(
7899
' ($
0 ( ;
$
$
<
=
+
1
.
.
,
+
1
%
$
1
-
.
!
0
"
!
=
$
0*
$
+
. /
+
1
3 - 4/
(
-
. /
.
+2
%
$
#
&1
(
-
#
3
5
'1
2
!
-
%
.
6 $
)
.
+
+
'
"
*.
'
7
-
1
.
)
8
9 .
!
'
0
8
7
.
:'
*
+
)
'
*
$
"
)
/
"%
)
=
! "
0
! "
1
>
! " &"
"6
$
"'
%*
'
$
&
*
+ 2
#
/
-
#-
-
)
&
)
)
'1 %
7 )
1
* 2 )
)
*
1
'
'
/
!
+
7 )
9 1
.
+
?,
>
$
:
7
.
(
@1
'
1
.
44
)
1
#
"
)
*
!
-
&
.
'
?*
?*
&
.
'/@
6
'
//@ '
'-
#
&
! .
1
'
!
.0
!
'
!
'
!
(
.>
.0
A
!
'
%
,
. 4
'
!
,
$
& 1
*
3
$
! .
0
0
! "
& .
&
6
3
-
. 4
-
! 0
()
.
.- /
3
(
#
&
#
(
+
"
! "
&&
$
&' *
%
#
#-
!
-
&* 2
1
#-
-
$ /
;
$
1
(
$
)
4
,
(/ .
+ 3
=
!
'
1-. 7BB '
%
<
&9 '
#-
( )'
$
-
#
-
'
+ 2
&: 3
+ 2
&
*
/
&#*
0
$
"
/
%
$
$
&
/
&
'
4.0
.0
)
& 1
)
&
'
1
/
7 )
.
& &!
& ' >
$ 8 7 )
& *
1 ) 3
"
&
.
5
%
.>
1
7
1
*
'
0
)
$ ,
7 )
&
&
! .
/
!
'
-
)
"
#
/
)
*
. 4
!
#
#
3
-
%
& 9 '
! "
& :%
$
"
&
4
/
-
.
+
&
)
. 4
' 3
.
(
.- /
' 1
- .$"
$ 44
.
C3
. 4
#
+2
3 - 4/
+
' 6 )
/
4
)
/
%
/
-
-
"
'
) ,
*
'& +
*
-
3
'
!
'
'
) *
-
'
#-
'' !
.
%
'* 2
(
'9 $
(
%- D
.0
.
?
E .
-
-
&
'
/
3 "4 3
4
.
4
! .
4
4
$
?4
-
': !
&
5" /
"
* 1
'
(
#
1
!
)
* 7
!
8
.
F
1
8
'
/
!
.
)
* +
7
$
$
#
.
*
*
-
'
)
* ,
'
'
! .
%
6
.
!
*&
$
@
+
$
-
*'
(! ?
**
-
*
) @
,
# !
1
!
)
-
*9 / +
,
-
# !
*:
*
)
-
!
$
'
$
6
)
-
#
*
/ >
/
$
44
*
"
)
'
/
!
.
*
*
! )
#
*
3
*
*
*
.
* 6
+
)
#
* &2
*
.
,
-
*
*
(
* '1 /
#
&
*
*
1 0
6
%
1
'
6 )
!
!
%
*
/
* *'
/
-
!
*
* 9#
(
3
-
!
. 4
)
#
%
* :.
#
!
*
)
)
'
*
$
'
*
.% -
1
*
1
.
$
-
-
.
*
0
*
-
,
!
'
7 )
!
7
.
.
$
*
)
'
)
,
2
.0
* & 2
!
#-
!
#-
#
+
/
/
/
1
1
" 1
%
'
#
$
-
#
)
'
-
!
$
'
-
)
$
*
%
'
* '
#
/
/
* *(
.0
'
'
0 *
7 )
)
1
./ $
* 9/
.
&
-
(
#
)
* :
)
*
#
3
)
/
A
1 )
$ *
-
)
"/
7
0
"
+
%
*
*
+
1
!
)
"-
.
+
9
!
'
!
.!
.
!
(
$
#
,
9
$ 44
$
9
.
.
$
)
$
94 #
6 )
/
95 3
#
,
$ 44
#
$ 44
/
+
!
.
)
)
C
,
$
'
$
$ 44
#
$
+
96 %
%
.
, 7 )
1
)
97 1 !
7 )
)
99 ,
)
,
?
!
.
5
+
"
9: #
.
/
9
.
'
5
+
"
0
%
44
0
. 4
A
!
9
G
*
6
%
9
!
)$ %
"
3
$
!
)
/
'
!
,
1
3
+
'
9
! .
$
+
9
'
.0
'
*
.
6
! .
.0
A
,
%
" 2
'
-