The learning parameter of the model is designed to capture actual strategies
that people adopt in cultural learning. Above in sections 2.5 and 8.5 we have
already discussed why and how people select models from whom they learn.
Without going into details, we can note that using such strategies is necessary
because we cannot check the truth of each piece of information we learn from
others. Thus we have to resort to a number of heuristics:first, we tend to learn
from our family and in-group, reflected by the assumption of the conversion
of households as units; second, we can follow the majority, which is imple-
mented by the“probabilistic”strategy in the model; andfinally, we can learn
from people of high prestige or people who give strong signals of their
commitment to their message, which is implemented by the parameters as
explained below.
The remaining two parameters are related to the work of the“apostles”in
the model. The“apostles”are represented by turtles that move freely in the
model world, changing their location to the next patch (household) on
every day, while changing directions occasionally. In terms of conversion,
they behave exactly as households: when a household decides whether
to convert, an apostle on a neighboring patch counts just like a converted
neighbor.^5 However, the apostles’effect on the households can be greater
than the effect of the neighboring households. This is taken into con-
sideration by the adding the“prestige-bias”parameter, the value of which
increases the probability of conversion if the household communicates with
an apostle. If the conversion-rate andprestige-bias add up to a probability
larger than one, the probability of conversion is always adjusted to one. This
makes sense in the real world: if a message catches on every time anyone
tries to pass it on, it does not matter if some of the messengers are special,
charismatic persons. In the model, two different learning heuristics are
modeled by a single prestige-bias parameter: prestige bias proper (copying
individuals with high status) and commitment signaling (copying individ-
uals who give costly signals of being committed to their message). Arguably,
the apostles’status was considered to be high onlybyinsiders;theirspecial
impact on outsiders (to be converted) could be based on their charismatic
presence and strong signals of commitment (so-called credibility-enhancing
displays, see section 8.5), such as enduring dangers and leading ascetic lives.^6
(^5) It would be intuitive to take into account when an apostle is exactly on the patch repre-
senting a household. However, the purpose of the model is to separate the question of mobility
(as the main characteristic of itinerancy) from other aspects of apostolic mission, which are
captured by a separate parameter (see section 9.4).
(^6) The code of the model is shown in Figure 9.4 in the Appendix to this chapter, complete with
comments in italics. Within the limits of this chapter, we cannot go into the technical details of
the code. However, the discussion of the model below will be self-explanatory and can be
followed without necessarily understanding all details of the code.
Social Networks and Computer Models 195