Page 194 - MODES of EXPLANATION

Basic HTML Version

changed as to how the gene codes for protein; the material happenings are unaltered. It is just
that as the now more common condition, the mutant is now considered the normal
advantageous character. The overwhelmed, previously dominant, “normal” gene still codes
for what it always coded. Dominant and recessive are not the territory, they are part of the
map for dealing with genetics. The model can be easily confused with material reality if
observer decisions are not recognized and acknowledged.
The above example shows that normalcy is an observer decision; it took Tainter and
Lucas (1983) to illustrate that significance is a material distinction, but it is in fact a decision,
just like dominance. Tainter and Lucas (1983) were explicit that the significance of an
archeological site does not have a material basis. A site may become significant because of a
new technology at hand, such as carbon dating or the ability to recognize the origins of
certain carbon fragments. A majority of problems that are challenging in science are not
material issues and so are not data problems.
The same error arises in the conviction that the critical experiment that will settle the
big issue is not just around the corner. That is rarely the case. The solution lies more often in
clear thinking and clever conception of logical types. For instance, there is an abundance of
experimentation on whether or not a higher temperature increases the rate of decomposition
of recalcitrant carbon in soils (Wixon, 2011). Unfortunately, each experiment is so locally
focused in time and space that non-linearities make the results appear to be at odds with
others, when in fact it is simply that the conflicting experiments are measuring something
different from some other logical type. Disagreement comes from local experiments over-
reaching the local universe of discourse in which they are set. All of the above conflicts arise
from some version of mistaking the model for the thing being modeled.
Models are important because their internal consistency gives the modeler a well-
founded confidence. Nevertheless, the price paid for the focus of models is the narrow scope
over which they apply. Models have to be quite local if they are to be consistent, otherwise
the model would be pulled off its parameters (its implied equation would break down). For
instance, one can model the shock wave of a terrorist’s bomb in a bus, but one cannot model
the larger issue that a terrorist for one side in a conflict is a freedom fighter for the other.
What to call the actor comes from a wider discourse within which the model for the bomb is
set.
Narratives Relative to Models: Making the Science Operational
The four pairs of contrasts
Given how easy it is to mistake the model for the thing being modeled, it is useful to identify
modeling relative to narrative, and the role of the observer in each. Zellmer et al. (2006)
developed a scheme that linked together four distinctions relevant to models:
Pattee’s (1978) distinction between laws and rules.
Allen and Hoekstra’s (1992) distinction between scale and type.
3