https://doi.org/10.31449/inf.v43i2.1888 Informatica 43 (2019) 209–219 209
Agent-Based Modeling of Society Resistance against Unpopular Norms
Kashif Zia, Dinesh Kumar Saini and Arshad Muhammad
Faculty of Computing and Information Technology, Sohar University, Sohar, Oman
E-mail: kzia@su.edu.om, dinesh@su.edu.om, amuhammad@su.edu.om
Umar Farooq
University of Science and Technology Bannu, Khyber Pakhtunkhwa, Pakistan
E-mail: umar@ustb.edu.pk
Keywords: agent-based modeling and simulation, unpopular norms, emperor’s dilemma, norm aversion
Received: June 24, 2017
People live in a society adhering to different types of norms, and some of these norms are unpopular.
This paper proposes an agent-based model for unpopular norm aversion. The proposed model is simulated
asking important “what-if” questions to elaborate on the conditions and reasons behind the emergence,
spreading and aversion of unpopular norms. Such conditions can thus be analyzed and mapped onto the
behavioral progression of real people and patterns of their interactions to achieve improved societal traits
particularly using the new social landscape dominated by digital content and social networking. Hence, it
can be argued that careful amalgamation of social media content can not only educate the people but also
help them in an aversion of undesirable behaviors such as retention and spreading of unpopular norms.
Simulation results revealed that to achieve a dominant norm aversion, an agent population must incorpo-
rate a rational model, besides, active participation of agents in averting unpopular norms.
Povzetek: Razvit je agentni sistem obnašanja množic, ki omogoˇ ca analizo odpora proti nezaželenim nor-
mam.
1 Introduction
Social norms are concepts and practices prevalent in a soci-
ety [7]. Formally, “Norms are practical prescriptions, per-
missions, or prohibitions, accepted by members of particu-
lar groups, organizations, or societies, and capable of guid-
ing the actions of those individuals” [21].
Norms are accepted which means that their existence is
evident from empirical inquiry. However, there is a con-
tradiction in the viewpoint of the notion of existence. One
view of norms existence (acceptance) is internalized that
incorporates it into individuals’ identity [21, 1]. The other
view uses the intentions of individuals’ as the criterion for
norm acceptance instead of the identity of individuals [21].
Therefore, conforming/accepting a norm corresponds to
the first view while following a norm relates to the sec-
ond view [3]. This distinction allows thinking of a norm
even without accepting it [21]. Norms describe a collective
behavior of groups, organizations, or societies but they are
the collective outcome of individuals’ cognition.
Norms have the power to transform into actions. This
can lead to norm transformation as well. Brennan et al
[3] have distinguished between conforming and comply-
ing (following) with norms. Similarly, they differentiated
between avoiding and acting opposite to norms. These ac-
tions are norm guided and in the absence of a norm, an
action would not be performed or it would be performed
but not in a similar fashion [21].
Norms play an important role in the development of so-
cial order [30]. They can change, create and affect behav-
iors. On the other hand, behaviors are capable of chang-
ing, creating and affecting norms [15]. Individual behavior
affects the behavior of other individuals in its range of in-
fluence [8]. The process is often defined as norm being
“externalized”. These externalities are able to reproduce a
regulatory impact on individuals’ behavior [12]. According
to Christine Horne, a higher degree of norm enforcement
have large sanctioning benefits [9]. She designed a norm
enforcement theory with the following features. In the case
of group welfare, sanctioning benefits have a positive effect
on norm and metanorm enforcement. However, the sanc-
tioning cost has a negative effect on norm enforcement. In
the case of social relations, interdependence has a positive
effect on norm enforcement. Similarly, sanctioning cost
and interdependence have a positive effect on metanorm
enforcement. Meta-norms are a particular type of norms
that regulate enforcement. Interdependence means the ex-
tent to which individuals value their relations. The experi-
mental analysis of the theory of enforcement has revealed
that theories that do not consider social relational contact
may produce faulty predictions.
Generally, an individual in a society is expected to be-
have according to societal norms. However, the equation is
not that simple. Following a societal norm does not mean
that an individual is accepting it. There may be a num-
ber of conditions and incentives that force an individual
to follow a social norm [21]. This clearly distinguishes

210 Informatica 43 (2019) 209–219 K. Zia et al.
the distinction between following and conforming to the
norm. If a norm is not confirmed or accepted from the
inside of an individual, just following it as a visible trait
is of weaker intensity. Hence, in the case of an individ-
ual, a norm can be followed as a result of social pressure,
but not accepted, if the individual’s personal belief does
not correspond to it. Contrarily, an individual may ac-
cept/conform to a norm, if personal belief corresponds to
following it. Christine Horne [10] has emphasized on re-
lationships, which are more important than individual per-
ception about norms. She argues that these relationships
can even persuade an individual to enforce a norm, even if
there is no apparent benefit of doing it. This situation be-
comes interesting when a particular norm is unpopular in
nature.
Unpopular social norms are those norms with which the
majority of people do not agree or believe in it internally. In
fact, people personally do not agree with unpopular norms
but still stick to them. Individuals’ may even unintention-
ally enforce others to follow them. Such cases in sociology
are dealt through a dilemma called, Emperor’s Dilemma,
as illustrated by Nkomo in [24]. Emperor’s dilemma re-
lates to a tale in which everyone shows fake admiration for
a new gown worn by an emperor even though the emperor
was naked. The cunning gown designers announced that
the (non-existent) gown would not be visible to those who
are not loyal to the emperor or who are really dumb. The
fear of being punished and of being identified as having in-
ferior societal traits, no one spoke the truth. The truth that
the emperor was in fact naked.
It is evident that the Emperor’s Dilemma is demonstrated
in many places around the world in one way or the other.
Whether it is foot-binding in neo-Confucian China or inter-
cousin marriages and dowry in Asia (indicated by Blake
in [2] and Hughes in [11], respectively), the nature of the
thought process is the same. People do not reveal what they
really believe from the fear of being identified as ignorant
or anti-social.
It is not that harmful if unpopular norms are followed
at an individual level. However, when a large population
adopts unpopular norms, following it becomes a kind of
default behavior that might influence the neutral part of the
population. As a consequence, it has been observed that
people even start enforcing unpopular norm which they dis-
approve in private. This behavior is generally termed as
false enforcement. Willer et al. in [29] focused on finding
out the reasons for false enforcement. According to them,
people falsely enforce unpopular norms to create an illu-
sion of sincerity rather than conviction. They performed
experiments using two scenarios, namely, wine tasting and
text evaluation. Experimental results revealed that people
who enforced a norm even against their actual belief, in
fact, criticized different alternate variations of an unpop-
ular norm. In short, their outcomes indicate that how so-
cial pressure can lead to false enforcement of an unpopular
norm.
Un-popular Norms (UNs) could have an adverse impact
on society and it is, therefore, sometimes necessary to op-
pose and possibly avert them. To achieve this goal, it is im-
portant to know the conditions which enable the persistence
of unpopular norms and models that support possible aver-
sion of these norms. This study attempts to elaborate on
the conditions and reasons behind the emergence, spread-
ing and aversion of unpopular norms in a society, using a
theory-driven agent-based simulation.
The rest of the paper is structured as follows. Section 1
introduced the research work presented in this work. Re-
lated work is provided in section 2. The current models
and then the proposed model is presented in section 3. De-
tailed analysis and comparison are provided in 4. Section
5 ends this paper with conclusions and future direction of
this work.
2 Related work
The propagation and transformation of norms co-evolve
with each other. Norms propagate through diffused influ-
ence. Since the subjects being influenced may have their
own perspective, they may decide to adhere or reject it.
As a result, the reciprocating influence of the subjects may
transform the norm itself. According to Macy and Flache,
exploration of scenarios of such a nature has been a sub-
ject of complex adaptive systems and they are investigated
by developing agent-based models [18]. Understanding the
emergence of norms in a society of agents is a challenge
and an area of ongoing research [27].
Studying norms in society have been one of the research
focus of agent-based modeling community. Theoretical
studies on norms such as those conducted by Conte and
Castelfranchi [6] and Meneguzzi et al. [20] explored that
agent are supposed to comply with social norms. The
sense of punishment from the society is evident as the pre-
dominant factor behind compliance of norms [4]. Stud-
ies conducted by Sanchez-Anguix et al. [25] and Sato and
Hashimoto [26] focused on the emergence of norms and
they described strategies showing how norms prevail in a
society. This is basically governed by societal influence.
Agents set their goals and frequently change their behav-
ior based on societal influence until a global equilibrium in
achieved [27]. Though lots of work has been done on the
emergence and prevalence of norms, very limited is car-
ried out for the aversion of unpopular norms. To the best,
our knowledge, our previous work [31, 22, 23] is the only
agent-based study on this exciting research area. Willer et
al. have pointed out many “empirical cases in which in-
dividuals are persuaded to publicly support behaviors or
beliefs that they privately question” [29]. The term, Prefer-
ence falsification, coined by Kuran [17] is defined as “the
act of misrepresenting one’s genius wants under perceived
social pressures”. According to him, an equilibrium is the
sum of three utilities namely, intrinsic, expressive, and rep-
utation. The intrinsic utility is about an individual’s per-
sonal satisfaction being part of the society. The expressive

Agent-Based Modeling of Society Resistance. . . Informatica 43 (2019) 209–219 211
utility is about an individual gain in the response of present-
ing himself/herself to be what is expected. The utility that
is acquired through the reaction of others is termed as rep-
utation utility. The concept of an unpopular norm is very
close to the concept of preference falsification, in which
individuals publicly lie about their privately held prefer-
ences [16]. According to Makowsky and Rubin [19], such
societies are “prone to cascades of preference revelation if
preferences are interconnected - where individuals derive
utility from conforming to the actions of others”. Further,
“ICTs and preference falsification complement each other
in the production of revolutionary activity. The former fa-
cilitates the transmission of shock while the latter increases
the magnitude of change that arises after a shock.” The util-
ity acts in two different ways in the propagation of unpop-
ular norms. At one end, it can force an individual to fol-
low an unpopular norm, or even falsely enforce it. On the
other end, it can propagate an opposite sentiment as a re-
sult of private preference revelation. There is a number of
evidence that a minority of activists (capable of revealing
their private preferences on will) can make a big difference
but in a conducive environment [13]. So, the relevant ques-
tion in this context becomes “Can a minority of activists
change an unpopular norm adopted by the majority?”.
3 The proposed extended model
To avert UNs, it is important to understand conditions that
might help to stop the propagation of these norms. Partic-
ularly, it is imperative to find the conditions necessary to
establish an alternative norm - a reciprocal norm of pre-
vailing UN, and the conditions that enforce people other
than activists to follow the alternate norm. This section first
introduces the social interaction model for following UNs
proposed by Centola et al. [5]. It, then, provides briefly our
previous extension to this model followed by the proposed
extension in this paper.
3.1 Centola’s model of norm aversion
Centola’s model [5] is capable of elaborating the conditions
and reasons behind the emergence, spreading, and aver-
sion of UNs in society but using theory-driven approach.
Consider an Un-popular Norm (UN) prevailing in a soci-
ety. Assume that a minority of the population truly be-
lieve in it due to some vested interest. Agents represent-
ing this population are termed as True Believerss (TBs).
Contrarily, a majority of the population do not believe in
the UN. Agents representing this population are termed as
Dis-Believerss (DBs). Figure 1(a) illustrates a sample dis-
tribution scenario.
Centola’s model is based on four variables explained be-
low:
1) Belief: an agent’s belief in UN which is 1 in case of
TBs and 1 in case of DBs.
2) Compliance: means that an agent is complying with
a UN or not? Initially, all TBs are complying
(compliance = 1) and DBs are not complying
(compliance = 1).
3) Enforcement: is an agent influence on the neighbor-
hood. Starting with a default value of 0, it can either
be 1 or 1.
4) Strength: is an agent’s resistance against compliance
of a UN.
An agent i’s belief is a static value. The value of compli-
ance may change using Equation 1.
compliance
i
=
  belief
i
if (
 beliefi
Ni
 NE
i
)>strength
i
belief
i
otherwise
(1)
enforcement
i
=
8
<
:
 belief
i
; if (
 beliefi
Ni
 NE
i
)> (strength
i
+k)
V
(belief
i
6=compliance
i
)
belief
i
; if (strength
i
 enforcement_need
i
>k)
V
(belief
i
=compliance
i
)
0; otherwise
(2)
Where,NE
i
= count of (Moore’s) neighbors enforcing
opposite belief and N
i
= count of (Moore’s) neighbors.
This means that an agent’s decision to comply with UN or
not is dependent on the enforcement of opposite belief by
the neighborhood. If NE
i
is greater than the strength of
a DB, the agent would comply against its belief. Since,
TBs compliance (which equals their belief about a UN)
and strength are already equal to 1, Equation 1 would not
change the compliance value of TBs.
When compliance is decided, an enforcement decision is
made next. Enforcement value may change using Equation
2.
Equation 3 is used to compute enforcement_need
i
- that is the need of enforcement reflecting influence of
neighborhood compliance.
enforcement_need
i
=
(1 beliefi
Ni
) NC
i
2
(3)
Where, NC
i
= number of (Moore’s) neighbors whose

212 Informatica 43 (2019) 209–219 K. Zia et al.
Figure 1: (a) Simulation set-up for 100 agents including 10% TBs. Initial values: TBs (belief = 1, strength = 1.0,
compliance = 1, enforcement = 0), DBs (belief = 1, strength = [0.01-0.29], compliance = 1, enforcement = 0). (b)
Changes in arrow directions in response of the application of Centola’s model.
compliance is different than the agent’s belief.
When an agent’s belief is equal to compliance (true is the
case of TBs and starting values of DBs), then the enforce-
ment will be equal to belief but only when strength en-
forcement_need of an agent is greater than a threshold vari-
ablek. Otherwise, it would remain 0. Since, the strength
of DBs is kept very low, thus the condition would not result
in enforcing 1 value by DBs. This condition will always
enforce a value of 1 by TBs. On the other hand, when an
agent’s belief is not equal to compliance (true is the case
of DBs with belief 1 and compliance 1 when Equation 1
is applied), the enforcement will be equal to the negation
of belief but only when the enforcement of opposite belief
in the neighborhood is greater than strength plusk value of
the agent. This means that such a DB itself start enforcing
a UN.
For example, the TB with ID 6 (see Figure 1 (b) - middle)
uses Equation 3 to calculate the enforcement_need value
equal to 1.6 for belief
i
= 1, N
i
= 5 and NC
i
= 4.
Thus, the value of compliance for the agent (from Equa-
tion 1) remains equal to its belief, because, neighborhood
enforcement ( (1)=5 0 , where 0 =NE
i
) is not greater
than its strength 1. However, the second condition of Equa-
tion 2 changes enforcement from 0 equal to 1, because, the
strength value of 1 multiplied with enforcement_need value
of 1.6 gives a much greater than the enforcement threshold
k which is considered 0.2 in this case. The same explana-
tion applies to TB named 2.
DB (BD check it please), numbered 55 (see Figure 1(b))
applies Equation 3 to get the enforcement_need value equal
to 1.33 forbelief
i
= 1,N
i
= 3 andNC
i
= 2. In this
case, the value of compliance for the agent (from Equation
1) changes to the opposite of its belief, because, neighbor-
hood enforcement ( ( 1)=3 2 , where 2 = NE
i
) is
greater than its strength value of 0.14. The first condition of
expression 2 changes the enforcement value from 0 equal
to 1 as neighborhood enforcement ( ( 1)=3 2 , where
2 =NE
i
) is much greater than the enforcement threshold
k (0.2 in this case) plus strength (0.14).
There are some DBs that do not comply at this point.
For example, DB 34 by using Equation 3 calculates the en-
forcement_need value equal to 1.2 forbelief
i
= 1,N
i
=
5 and NC
i
= 2. The value of compliance for the agent
(from Equation 1) remains unchanged, because, the neigh-
borhood enforcement ( ( 1)=5 1 , where 1 = NE
i
)
is not greater than its strength value of 0.216. The second
condition of Equation 2 would make enforcement from 0
equal to 1, because strength (0.216) multiplied with en-
forcement_need (1.2) is slightly greater than the enforce-
ment threshold k considered 0.2 in this case. The same
applies to DB 10 and 42. Contrarily, the enforcement of
DB 83 remains 0. These DBs, however, start complying
at next iteration (see Figure 1 (b) - right) due to combined
enforcement of their neighbors.
3.2 Our previous extension
Since, acpdb compliance in basic centolla’s model is un-
desirable, in our previous work [31], we extended it and
introduced a special kind of DBs (called Activists (ACTs))
with more desire to avert (act against) a UN. These ACTs
are triggered by the presence of TBs in the surrounding,
particularly who are enforcing. Their strength is progres-
sively incremented proportionally to the intensity of en-
forcement from TBs. The strength of an ACT is calculated
using Equation 4.
strength
i
=strength
i
+ (
E
jb
N
i
) (4)
Where,E
jb
= is the number of enforcing TBs.

Agent-Based Modeling of Society Resistance. . . Informatica 43 (2019) 209–219 213
Figure 2: Simulation results of the basic Centolla’s model for various scenarios based on number of agents (considered
50, 100, and 200) and threshold value k - showing an agent’s desire to comply (considered 0.2, 0.5, and 0.8).
Figure 3: Simulation results of our previous extension to Centolla’s model for various scenarios based on number of
agents (considered 50, 100, and 200) and threshold value k - showing an agent’s desire to comply (considered 0.2, 0.5,
and 0.8).
3.3 The proposed extension
In this paper, the model is further extended to incorporate
the decision-making of a DB as a result of neighborhood
condition. It is proposed that DBs (who are not ACTs)
should not be considered as entirely a numb entity. We
propose a decision-making model represented in Equation
5. In this model, the strength of DBs (who are not ACTs)
is changed (increased or decreased) based on its type be-
ing either “optimistic” or “pessimistic”. The difference be-
tween percentage of enforcing TBs (termed as, P
jb
) and
percentage of complying DBs (termed as, P
jd
) is divided
by neighborhood density (N
i
) times the fraction of DBs of
that type (consideropt for an optimistic and \1 opt
00
for a
pessimistic DB). If an agent belongs to the optimistic cat-
egory, its strength would be increased/decreased based on
the difference of “true enforcement” (represented as P
jb
)
and “false compliance” (represented as P
jd
). When fast
compliance is more then the strength will decrease. On
the other hand, when true enforcement is more then the
strength will increase.
4 Evaluation and results
4.1 Simulation environment
NetLogo [28] - a popular agent-based simulation tool with
support for grid-based spaces, is used to simulate the work
presented in this work. The agents reside on cells of a
spatial grid. We have used the concept of Moore’s neigh-
borhood to represent the surrounding of an agent - a very

214 Informatica 43 (2019) 209–219 K. Zia et al.
strength
i
=
 strength
i
+ (P
jb
 P
jd
)=(N
i
 opt); if iisoptimistic
strength
i
+ (P
jb
 P
jd
)=(N
i
 (1 opt)); otherwise
(5)
Figure 4: Simulation results of the proposed extension (with 10% agents of total population being optimistic) to Centolla’s
model for various scenarios based on number of agents (considered 50, 100, and 200) and threshold value k - showing an
agent’s desire to comply (considered 0.2, 0.5, and 0.8).
popular strategy in many cell-based spatial configurations
[14]. For a coarse-grained evaluation, we used a simulation
space consisting of a torus of 17 17 cells. Figure. 1(a)
provides an illustration of this space filled with 100 agents.
4.2 Results and discussion
4.2.1 Previous findings
Due to the spatial nature of the neighborhood, it was ex-
pected that a more dense population is susceptible to more
DBs compliance. This fact is evident from the results
shown in Figure. 2. Further, DBs’s compliance is inversely
proportional to the value of k - an agent’s desire to comply.
Ironically, in all cases depicted in Figure. 2, the population
achieves stability always being attracted towards various
fixed points.
In our previous work [31], it was observed that in highly
dense conditions with a large number of norm aversion
ACTs, the aversion of unpopular norms can be achieved.
This fact is highlighted in Figure. 3. There is a striking
similarity between the basic model and our previously ex-
tended model whose results are presented in Figure. 2 and 3
in corresponding order. It is learned that the cases com-
prise of smaller values of k and a large number of agents
are worst than the rest of the cases. A marginal improve-
ment was achieved by introducing the ACTs where com-
paratively less number of DBs were witnessed complying
with a UN.
4.2.2 Current findings: a brief analysis
This model uses optimistic DBs that are intrinsically be-
lieving in averting the UN. Simulation work conducted in
this paper uses three different numbers of these optimistic
DBs, which are counted as 10, 20 and 30% of the total
population. It was learned that the proposed model signif-
icantly reduces the number of DBs complying with a UN.
Even the scenario considered as a worst one (the one com-
prises of 200 agents and a threshold value k = 0.2) achieved
a 100% improvement by drooping compliance rate from
70% to 35%. This is illustrated in Figure. 3 and 4.
When the proposed model is compared with the previous
model, it was noted that DBs’s compliance comparatively
gets worse as the number of agents’ increases irrespective
of the value of k.
The cases where the number of agents is 200 always
perform worse than other cases (comprising of 50 or 100
agents). This can be noticed while comparing the results
presented in Figure. 3 with 4). Overall, with an increase in
the number of optimistic DBs, the results get improved as
witnessed by comparing the results given in Figure. 4, 5,
and 6).
4.2.3 Current findings: a detailed analysis and
comparison
In this section, we present a detailed analysis of the simu-
lation results. The simulation space for these experiments
comprises a torus of 33 33 cells. 1000 agents are placed
on cells without overlapping. Figure 7(a) provides this

Agent-Based Modeling of Society Resistance. . . Informatica 43 (2019) 209–219 215
Figure 5: Simulation results of the proposed extension (with 20% agents of total population being optimistic) to Centolla’s
model for various scenarios based on number of agents (considered 50, 100, and 200) and threshold value k - showing an
agent’s desire to comply (considered 0.2, 0.5, and 0.8).
Figure 6: Simulation results of the proposed extension (with 30% agents of total population being optimistic) to Centolla’s
model for various scenarios based on number of agents (considered 50, 100, and 200) and threshold value k - showing an
agent’s desire to comply (considered 0.2, 0.5, and 0.8).
setup.
Simulation results are analyzed based on the following
four quantities:
1) DBComplBCount: is the number of Dis-Believerss
which comply with the Un-popular Norm, B, against
their belief.
2) DBFollBCount: is the number of DBs which do not
comply with the UN, B, but follow it against their be-
lief.
3) DBComplACount: is the number of DBs which com-
ply with the alternate norm, A, but still do not believe
in it.
4) DBBelACount: is the number of DBs which comply
with the alternate norm, A, and believe in it.
The purpose and intention of the proposed model are to
reduce the value of DBFollBCount because these agents
are unsure and their belief can potentially be averted. The
possible aversion may transform agents status from “fol-
lowing” to those which are “complying” with the alternate
norm (DBComplACount).

216 Informatica 43 (2019) 209–219 K. Zia et al.
Figure 7: NetLogo Simulation. (a) Setup of 1000 agents with 5% TBs and 5% ACTs. TBs, ACTs, and DBs are represented
as blue coloured triangles, blue coloured persons, and green coloured triangles correspondingly. (b) Results of basic
Centola model [5]. Equilibrium state, where all DBs now comply with the unpopular norm, B, against their belief.
The basic model proposed by Centola [5] formulates the
spread of a UN only. The results of the application of the
model settle in an equilibrium state after the 5
th
iteration.
Figure 9(a) visualises the concept presented in Figure 7(b).
It is evident from the results presented in Figure 9(a) that all
DBs started with following the UN, quickly, started com-
plying with it.
After all, DBs started complying with the norm, B, a
change in strategy was tested. The ACTs was activated to
play their role as proposed in [31]. The extended model
proposed by Zareen et al. [31] reached at an equilibrium
between 10
th
to 12
th
iteration. Figure 9(b) visualises the
concept presented in Figure 8(a).
It is evident from the results shown in Figure 9(b) that
DBs started complying with the alternate norm, A, under
the influence of ACTs.
The number of DBs which transformed to compliance
state merely changed to the following state again. Start-
ing with an increase in the following agents, a decline was
observed, however, it did not drop to 0. DBs following
and complying to the norm, B, stabilizes with followers
more than agents which are complying. As shown in Fig-
ure 8(a), DBs in the neighborhood of ACTs started follow-
ing and complying norm, A, against their belief.
The proposed extended model achieved equilibrium with
promising results. Figure 10(a) visualises the scenario pre-
sented in Figure 8(b). It is evident from the results given
in Figure 10(a) that DBs started complying with alternate
norm, A, under the influence of ACTs. Further, the ma-
jority of them started complying norm, A, with a belief in
it. In response, the number of DBs following the norm, B,
reduced to almost nothing.
Finally, we further increased the number of TBs and
ACTs to a comparison with the scenarios just discussed.
It was learned from Figure 10(a) and 10(b) that the pat-
tern and state changes are similar. However, the aggregate
number of DBs in state DBBelACount and DBFollBCount
has decreased substantially when compared with the aggre-
gate count of DBComplACount and DBComplBCount. It
means that the DBs who believed in the norm, A, was de-
creased by almost 50% when the number of TBs and ACTs
were doubled.
4.2.4 Discussion
The main objective of the proposed model was to reduce
the number of disbelievers complying with an unpopular
norm, B. It is clear from the simulation results given in
Figure 9 that our previous model presented in [31] reduced
the number of complying agents to 25% of the whole pop-
ulation as compared with 95% obtained by the standard
model. However, 50% agents still follow the norm, B, and
only 20% start complying with the alternate norm, A. The
credit goes to the introduction of ACTs in the population of
agents.
The introduction of optimistic agents in our current ex-
tension proposed in this paper significantly improved these
results. Though the number of disbelievers complying with
an unpopular norm does not change much, however, the
majority of disbelievers started believing in alternate norm
instead of following an unpopular norm. This is evident
from comparing Figure 9(b) and Figure 10(a) with each
other.
5 Conclusion
It is argued that for societal good, it is necessary to oppose
and possibly avert unpopular norms. This work is an at-

Agent-Based Modeling of Society Resistance. . . Informatica 43 (2019) 209–219 217
Figure 8: NetLogo Simulation. (a) Setup of 1000 agents with 5% TBs and 5% ACTs. TBs, and ACTs are represented
as blue coloured triangles and blue coloured persons correspondingly.The rest of the agents are DBs.Simulation Result of
the extended model proposed by Zareen et al. [31]. Equilibrium state, where DBs in the neighborhood of ACTs started
following (blue) and complying (red) norm, A, against their belief. (b) Simulation results of the current proposed extended
model.
tempt to realise the conditions that result in the emergence
of unpopular norms and define situations under which
these norms can be changed and averted. It presented an
agent-based simulation for unpopular norm aversion. It
utilised the reciprocal nature of persistence and aversion of
norms to define situations under which these norms can be
changed and averted. The simulation results revealed that
in addition to agents actively participating in averting the
unpopular norm, incorporating a rational decision-making
model for normal agents is necessary to achieve a dominant
norm aversion. Further, it was learned that the inclusion of
true believers and activists play a significant role in norm
aversion dynamics.
In short, this study revealed that more educated and
socially active individuals are key to reduce undesirable
norms in society. The significance of this fact is also ap-
plicable to digital societies primarily created by social net-
working applications nowadays.
References
[1] Cristina Bicchieri, Erte Xiao, and Ryan Muldoon.
Trustworthiness is a social norm, but trusting is
not. Politics, Philosophy & Economics, 10(2):170–
187, 2011. URL:https://doi.org/10.1177/
1470594X10387260.
[2] C Fred Blake. Foot-binding in neo-confucian china
and the appropriation of female labor. Signs: Jour-
nal of Women in Culture and Society, 19(3):676–
712, 1994. URL:https://doi.org/10.1086/
494917.
[3] Geoffrey Brennan, Lina Eriksson, Robert E Goodin,
and Nicholas Southwood. Explaining norms. Ox-
ford University Press, 2013. URL:https://doi.
org/10.1017/s0266267114000467.
[4] Will Briggs and Diane Cook. Flexible social laws.
In International Joint Conference on Artificial Intelli-
gence, volume 14, pages 688–693. Citeseer, 1995.
[5] Damon Centola, Robb Willer, and Michael Macy.
The emperor’s dilemma: A computational model of
self-enforcing norms 1. American Journal of So-
ciology, 110(4):1009–1040, 2005. URL: https:
//doi.org/10.1086/427321.
[6] Rosaria Conte and Cristiano Castelfranchi. Are in-
centives good enough to achieve (info) social order?
In Social Order in Multiagent Systems, pages 45–61.
Springer, 2001. URL: https://doi.org/10.
1007/978-1-4615-1555-5_3.
[7] Robert EL Faris. Handbook of modern sociology.
1964.
[8] Michael Hechter and Karl-Dieter Opp. Social norms.
Russell Sage Foundation, 2001.
[9] Christine Horne. Explaining norm enforce-
ment. Rationality and Society, 19(2):139–170,
2007. URL: https://doi.org/10.1177/
1043463107077386.
[10] Christine Horne. The rewards of punishment: A rela-
tional theory of norm enforcement. Stanford Univer-
sity Press, 2009. URL: https://doi.org/10.
1017/s0003975609990208.

218 Informatica 43 (2019) 209–219 K. Zia et al.
Figure 9: Simulation outcome in terms of number of DBs in various states against time of: (a) the basic Centola’s model
[5]. (b) the extended model proposed by Zareen et al. [31].
Figure 10: For optimistic agents, the simulation outcome of the proposed extended model in terms of number of DBs in
various states against time for the scenarios comprises: (a) 5% TBs and 5% ACTs. (b) 10% TBs and 10% ACTs.
[11] Diane Owen Hughes. From brideprice to dowry
in mediterranean europe. Journal of family history,
3(3):262–296, 1978. URL: https://doi.org/
10.1177/036319907800300304.
[12] Coleman James. Foundations of social theory. Cam-
bridge, MA: Belknap, 1990.
[13] Habibul Haque Khondker. Role of the new me-
dia in the arab spring. Globalizations, 8(5):675–
679, 2011. URL:https://doi.org/10.1080/
14747731.2011.621287.
[14] Tobias Kretz and Michael Schreckenberg. Moore
and more and symmetry. In Pedestrian and evac-
uation dynamics 2005, pages 297–308. Springer,
2007. URL: https://doi.org/10.1007/
978-3-540-47064-9_26.
[15] Dorothea Kübler. On the regulation of social
norms. Journal of Law, Economics, and Organiza-
tion, 17(2):449–476, 2001.
[16] Timur Kuran. The inevitability of future revolu-
tionary surprises. American Journal of Sociology,
100(6):1528–1551, 1995. URL: https://doi.
org/10.1086/230671.
[17] Timur Kuran. Private truths, public lies: The so-
cial consequences of preference falsification. Har-
vard University Press, 1997. URL:https://doi.
org/10.1177/000169939603900109.
[18] Michael Macy and Andreas Flache. Social dy-
namics from the bottom up: Agent-based mod-
els of social interaction. The oxford hand-
book of analytical sociology, pages 245–268,
2009. URL: https://doi.org/10.1093/
oxfordhb/9780199215362.013.11.
[19] Michael D Makowsky and Jared Rubin. An
agent-based model of centralized institutions, so-
cial network technology, and revolution. PloS one,
8(11):e80380, 2013. URL: https://doi.org/
10.1371/journal.pone.0080380.
[20] Felipe Meneguzzi, Odinaldo Rodrigues, Nir Oren,
Wamberto W Vasconcelos, and Michael Luck. Bdi
reasoning with normative considerations. Engineer-
ing Applications of Artificial Intelligence, 43:127–
146, 2015. URL:https://doi.org/10.1016/
j.engappai.2015.04.011.

Agent-Based Modeling of Society Resistance. . . Informatica 43 (2019) 209–219 219
[21] Paul Morrow. The thesis of norm transformation in
the theory of mass atrocity. Genocide Studies and
Prevention: An International Journal, 9(1):8, 2015.
[22] Arshad Muhammad, Kashif Zia, and Dinesh Kumar
Saini. Agent-based simulation of socially-inspired
model of resistance against unpopular norms. In 10th
International Conference on Agents and Artificial In-
telligence, pages 133–139, 2018. URL: https:
//doi.org/10.5220/0006735501330139.
[23] Arshad Muhammad, Kashif Zia, and Dinesh Ku-
mar Saini. Population dynamics necessary to avert
unpopular norms. In Agents and Artificial Intelli-
gence, pages 64–75. Springer International Publish-
ing, 2019. URL:https://doi.org/10.1007/
978-3-030-05453-3_4.
[24] Stella M Nkomo. The emperor has no clothes:
Rewriting “race in organizations”. Academy of
Management Review, 17(3):487–513, 1992. URL:
https://doi.org/10.5465/amr.1992.
4281987.
[25] Victor Sanchez-Anguix, Vicente Julian, Vicente
Botti, and Ana García-Fornes. Tasks for agent-based
negotiation teams: Analysis, review, and challenges.
Engineering Applications of Artificial Intelligence,
26(10):2480–2494, 2013. URL: https://doi.
org/10.1016/j.engappai.2013.07.006.
[26] Takashi Sato and Takashi Hashimoto. Dynamic social
simulation with multi-agents having internal dynam-
ics. In New Frontiers in Artificial Intelligence, pages
237–251. Springer, 2007. URL: https://doi.
org/10.1007/978-3-540-71009-7_21.
[27] George A V ouros. The emergence of norms via con-
textual agreements in open societies. In Advances
in Social Computing and Multiagent Systems, pages
185–201. Springer, 2015.
[28] Uri Wilensky and William Rand. An introduc-
tion to agent-based modeling: modeling natural, so-
cial, and engineered complex systems with NetLogo.
MIT Press, 2015. URL:https://doi.org/10.
1063/pt.3.2884.
[29] Robb Willer, Ko Kuwabara, and Michael W Macy.
The false enforcement of unpopular norms. American
Journal of Sociology, 115(2):451–490, 2009. URL:
https://doi.org/10.1086/599250.
[30] H Peyton Young. Social norms. 2007.
[31] Zoofishan Zareen, Muzna Zafar, and Kashif Zia. Con-
ditions facilitating the aversion of unpopular norms:
An agent-based simulation study. International Jour-
nal of Advanced Computer Science and Applica-
tions, 7(7), 2016. URL:https://doi.org/10.
14569/ijacsa.2016.070769.

220 Informatica 43 (2019) 209–219 K. Zia et al.