Computer models have become part of the standard methodological reper-
toire of cognitive sciences and the use of computer models has accompanied
the cognitive turn in variousfields.
Without spending much time on the philosophy of modeling, it is appro-
priate to consider some basic theoretical points before we start to work on our
specific model. In a general sense, models can be thought about as represen-
tations of phenomena that capture essential features of the phenomenon in
question. For example, a map of a geographical location provides an abstract
representation of the area. The same is true even of a hand-drawn sketch of the
place. In both cases, the model captures some features while ignoring others.
Note that maps and sketches can be made for various purposes and thus the
features included can vary. In any case, the model is always selective and
simplifying, including only features that are relevant for some purpose, such as
distances, proportions, buildings, elevation, vegetation, demographic data, and
so on. Sometimes we create models that are mainly tools to think with and
sometimes we build them from actual data and expect them to produce
realistic outcomes.
The relationship between the model and the phenomenon it represents can
be of several kinds. For example, a model house is adownsized modelof a real
building. Classical mechanical models of celestial bodies in physics areideal-
ized modelsin the sense that they have only mass.Analogical modelsrecreate
some phenomenon in a different medium, such as computer models of
cognitive processes. Finally,phenomenological modelsreflect surface charac-
teristics without paying attention to structural features, such as a wind-up
mouse that rolls on wheels or a more complex robot version that can avoid
objects. Note that most actual models combine more of these approaches to
some degree. A crucial factor to be paid attention to when creating models is
the degree of detail and complexity that goes into them. We can always add
details to a model to make it phenomenologically more accurate, yet by doing
so we are often losing control over what is happening structurally below the
surface. In other words, if we make a model too complex, we might not be able to
find out why the model shows the real-life features that we wanted to understand
originally. Keeping a model too simple, however, might provide only trivial
results or insights that are only very loosely related to real-life phenomena. As
long as our purpose is to use the model to learn about real-life phenomena, there
is always a trade-off between realism and inferential potential.
We can sort computer models of religion into three broad categories. First,
we can model religious cognition, including the representation of knowledge
and decision-making. Examples in this category involve modeling decisions to
perform specific kinds of rituals (Biró, 2011), or the mental representation and
production of commission narratives (Czachesz, 2007a; Czachesz & Lisdorf,
2013). Second, the social structures and their changes in religious groups and
their expansion can be studied. Although these phenomena can be addressed
188 Cognitive Science and the New Testament