Page 192 - MODES of EXPLANATION

Basic HTML Version

Chapter 14
Narratives and Models in Complex Systems
Timothy F.H. Allen,
Edmond Ramly, Samantha Paulsen, Gregori Kanatzidis, and
Nathan Miller
Humans often deal with the world through narratives. Mechanistic scientists do not
necessarily recognize when they are using narratives, and some might imagine narratives as
non-scientific. The thesis in this chapter is that narratives are in fact the deliverable in
science, and that models are a means of improving the quality of that bottom line. It may
appear that we are asserting an anti-realist point of view, but that is not the case. Most of the
time, we are agnostic as to reality, while accepting that the writer, Allen, is in his material
study, typing on a real keyboard. That would be a soft realism, which is perfectly acceptable.
Our case against hard realism is that reality is often used prematurely as an intellectual crutch
in the scientific endeavor. We wish to tighten up standards by getting a clear view of
narrative and modeling.
Our reservations about realism come from the way it encourages slovenly thinking
and careless analysis, which muddles the discourse. All we do in science is associated with
levels of analysis that determine what is in, what is out, what is related and how, and so on.
We have no reason to suppose that there are levels of analysis in the real external world.
Inserting reality into experience introduces random elements at undefined levels of analysis.
Small changes in those levels of analysis are particularly sinister, because it is easy to
overlook them, even though the consequences for the investigation might be large. Levels of
analysis are subtle and need to be treated in a nuanced manner. Narratives are problematic for
modernist realists in science because there is no obvious way for a narrative to be tied to
reality in any straightforward manner.
Funtowicz and Ravetz (1992) make the case that all modernism has an external
reference for quality that pertains to external reality: the closer the model is to reality, the
higher its quality. That would be the default setting for the majority of biologists. Modernism
does not seem to acknowledge that we only have access to experience and data; we do not
have access to reality except through our senses. Even the apparently clean shot at reality
resulting from a methodically rigorous experiment is misleading, because embedded there is
a host of arbitrary, unstated decisions. Modernists view measurements as a refuge that is
concrete. Unfortunately, measurements are never clean, and are set in models full of arbitrary
decisions. Robert Rosen (2012: 214) said pointedly that the most abstract thing you can do is
to make a measurement, because you have to exclude all the rest of the universe to get it. As
a result, there is in fact much more intrusion from arbitrary decisions of the observer than
modernists usually realize. When the study is over, or at least at a stage for reflection, realism
does no harm and may hearten scientists for the tedium of data collection and analysis.
Changes in reference occur often in biology. There is frequently someone at the back
of the lecture room who says, “Yes, but what about such and such?” imagining it is validly
part of the discourse. In fact, the “such and such” is often in a different universe that causes it
1