I think what Mike has been describing can be thought of in terms of the
development of a research system. I'll write as if this in the singular, for
clarity, although my gut tells me it's not. Consider an extended triangle
that models the coding system, itself in development as the codings and
interpretations are in flux. As coding proceeds the subjects (coders) begin
to fall into agreement. the "system" of interpretation becomes more "stable"
(or annealled) as the interpretations reach increasing levels of
reliability, and the subjects, rules, artifacts, and most likely also the
division of labor and object co-develop. I pose the analysis in this way,
because there seems to be some purchase in thinking of methodology as work
focused primarily around the relationships of artifacts, rules, and object of
the research system, so that the first system is insensitive to "who" its
subjects (researchers) are. One might be tempted to pose a second system
here, whose object is the development of the first. My gut tells me that
modeling methodology as a "system" will be difficult, and perhaps that is
where the trouble lay.
A problem arises as the research (first) system is "perturbed" (I borrow the
sense from physics -- in bronfenbrennerian terms, this might be more akin to
an ecological transition) as a new coder replaces an old coder. All
relationships involving the new coder and the old coder and data must be
re-established. The interesting result is that there is relatively little
convergence between the new (perturbed) system and the prior system. Why?
Can CHAT explain this?
After the perturbation of the research system with the introduction of a new
coder, the system settles into a new configuration, distinctly different from
that of the prior. I posit that the research system is *sensitive* to the
composition of its subjects, in direct tension with the object of the
methodological work, which arguably is about making research results
invariant across coders. What I'll hope you'll notice about this
off-the-cuff analysis of Mikes situation, is that it does not describe the
situation in terms of truth or reality, but in terms of making
interpretation with high reliability, that is to say, with the same data, the
system converges semiotically in a manner insensitive to the composition of
the collective subject. It is about making repeatable semiotic formations in
the system, starting from the same data.
How? The path is not obvious to me at this point, although I'm interested in
the following: (1) this seems to be a problem that parallells what science
classrooms attempt to achieve, (2) there may be a worthwhile contrast to
consider how and why Peter S. uses "setting" instead of "system" and (3) in
light of boundary object work, is this methodological problem intractable?
bb
This archive was generated by hypermail 2b29 : Fri Aug 01 2003 - 01:00:08 PDT