The Individual and Society 407squares, and rooms. One can envision other geometries, how-
ever, to capture different communication structures (Nowak
et al., 1996). A one-dimensional geometry in which people
interact mainly with neighbors to their left and right corre-
sponds to a row of houses along a river or a village stretching
along a road. In this case, strong clustering occurs because of
well-pronounced local interactions between nearest neigh-
bors. Polarization, however, is inhibited because members of
the majority cannot encircle members of the minority and
overwhelm them. Far more elaborate geometries of social
space can also be envisioned. In the real world, many differ-
ent geometries no doubt co-occur and thus determine the
dynamics of social influence. The availability of telephones,
e-mail, and common areas for shopping and recreation
clearly add many dimensions to the effective geometry in
which interactions occur. The combined effects of such
geometries play a significant role in determining the form and
outcome of social influence.
A fourth critical factor represents the weight an individ-
ual attaches to his or her own opinion as compared to the
opinions of others. This variable, referred to as self-influence,
corresponds to psychological states like self-confidence,
strength of conviction, and belief certainty. An individual’s
self-influence is correlated with his or her strength, although
the absolute value of self-influence varies as a function of
topic or social setting. When an issue is new or confusing, for
example, self-influence is correspondingly lower, reflecting
the fact that no strong opinion has formed and everyone is
relatively open to external influence. When an issue is famil-
iar and personally important, however, self-influence attains
its maximum value for everyone, reflecting the greater im-
portance of one’s own opinion compared to others’ opinions.
Because issue familiarity is assumed to be the same for all
individuals in a given simulation, variation in self-influence
is a direct reflection of variation among individuals in their
respective strength.
The dynamics of social influence are determined by the
value of self-influence relative to the total influence of other
individuals. When self-influence is low, individuals may
switch their opinions several times during the course of sim-
ulations. This has the effect of destabilizing clusters. For
topics that are unfamiliar, then, one observes heightened
dynamics that promote unification based on the majority
opinion. However, if self-influence is greater than the com-
bined influence of others, dynamics tend to be dampened
altogether, unless sources of noise (random external factors)
are present. Because noise works jointly with social influ-
ence, noise-induced changes are typically in the same direc-
tion as majority influence. Introducing a random factor that
by itself would not favor any position can thus neutralize the
effect of self-influence and enhance the effect of majority
opinion. Very high values of noise, however, can dilute the
effects of social interaction as well, producing random
changes in opinion.Social Change and Societal TransitionsThis general approach to the modeling of social processes has
proven useful in generating insight into the dynamics of social
change, including major societal transformations (Nowak &
Lewenstein, 1996; Nowak, Lewenstein, & Szamrej, 1993;
Nowak & Vallacher, 2001). This approach successfully mod-
els social change when a source of bias is introduced that
makes the minority opinion more attractive than the majority
opinion. The results of simulations reveal that rapid social
change occurs in a manner that is remarkably similar to phase
transitions in physical phenomena. Expressed metaphori-
cally, changes enter as bubbles of new within the sea of old,
and social transitions occur as these bubbles expand and be-
come connected. Thus, for example, a new political ideology
or lifestyle fashion that resonates with existing values or in-
terests is introduced into a social system and is immediately
embraced by pockets of people in different areas. These pock-
ets become increasingly connected over time, until at some
point the new idea achieves widespread dominance over the
old idea.
Computer simulations also indicate, however, that the
bubbles of the old manage to stay entrenched in the sea of the
new. The strongest and best-supported individuals holding
the old position, moreover, are the most likely to survive
pressures associated with the new position. This, in turn,
means that the old position is likely to display a rebound
effect when the bias toward the new position disappears or
is somehow reversed. This scenario provides an explanation
for the return of leftist governments in Eastern Europe after
their overwhelming defeat in the elections in the late 1980s.
This model of societal transition stands in marked contrast
to the conventional view of social change, which holds that
individuals gradually switch from an old set of attitudes or
preferences to a new set of ideas. From that perspective, new
ideas spread more or less uniformly through a society at a
constant and relatively slow rate. The simulation model
allows for this mode of social change as long as the social
system is near a relatively stable equilibrium and noise is not
a significant factor in dictating the system’s dynamics
(Nowak et al., 1993). The incremental scenario, in other
words, may effectively characterize how change occurs in a
stable society (e.g., a gradual shift from liberalism to conser-
vatism or vice versa), but it does not capture the nature of
change defining periods of rapid social transition.