Hi Matt,
>> For the social scientist, the approach raises two problems:
>>
>> 1) Too much reflection means too much attention to models of the world.
>> To ask the right questions means having unbiased data on how people in
>> some context of interest actually behave.
>>     
> I take it that when you say context of interest you are inferring that
> this is a model of the world. 
By context I mean some particular domain of human behavior that is 
believed to operate independently enough from others to name and study 
it.  Societal impacts of science and technology would be an example of 
what I mean by context of interest.
> I understand you as meaning that
> context is unstable, always shifting, as a natural outcome of
> reflection.  The act of shifting contexts and perspectives and between
> models of the world is reflexivity. 
Reflection on one's own experiences and comparing them with others 
(reflexivity) won't necessarily result in correct conclusions about why 
people do the things they do, or their larger social implications.  
Scientific work is incremental and academic departments usually operate 
more or less in parallel with others.   So, by design there's a lot of 
correlated work (and I'd imagine thought too).   Of course, focus can be 
good for punching through relevant problems in specific contexts..

To the point, it raises doubts in my mind just to what extent we can 
treat subjective reports of scientists and technologists as independent 
samples.  
> I'm, however,
> unclear on the relationship of unbiased data to the framework you are
> proposing.  
Suppose Bob's got an idea for an experiment and a paper to go with it.  
He runs the experiment and it fails to turn out the way he thought but 
reveals a better experiment which he also then runs and it results in an 
appealing outcome and insight.  Now Bob writes the paper with a new 
plausible sounding hypothesis that nicely yields to the outcome and 
conclusion (as if the original hypothesis and experiment never had 
occurred).   The paper is cited all over the place and Bob's a big hero. 

To understand problem solving in Bob's context, realizing their are 
potentially lots of Bobs, is it such a good idea to go on Bob's 
reflections and Bob's buddies?    Wouldn't it be better to devise a way 
to monitor Bob's actual day to day work in some minimally intrusive 
way?  One should worry about the accuracy of `reflections' and the 
reflexive cross comparison of them.   It strikes me that research 
results in this area are vulnerable to self-aggrandizing delusions 
shared by the researcher and researched (both of the scientist type).
>> 2) It's typically not possible to sufficiently influence or observe
>> people to understand cause and effect across individuals or groups.
>> The insights gained from reflexive participation will just be the kind
>> of models we get living life (but with fancied-up language to sound more
>> important than they are).  Seems to me this kind of modeling is more the
>> domain of the intelligence agencies than universities.
>>     
>
> I take it that when you say that there is an impossibility to
> influence or observe then you are speaking from a particular model of
> the world.  I cannot understand what you mean by sufficiency until I
> better understand where you are coming from. 
I don't think a cybernetic / control system approach to understanding 
human behavior is impossible, just expensive and something only certain 
governments could sustain in general form.   One might imagine that..

1) participants have models which may change in accordance to new 
observables
2) the models are shared to some extent (either to communicate or 
manipulate)
3) the participants are autonomous
4) the participants all have something at stake -- most aren't faking it
5) ..but some aren't what they seem -- they are there only to perturb 
and measure

You can imagine that in this kind of scenario, you'll find individuals 
acting in authentic, motivated ways.
If a set of participants in this situation had large (but invisible) 
cash resources to draw on, and were willing to tolerate risk (e.g. 
spies), they could in some sense facilitate the kind of data collection 
that would be needed to truly inform the agents in an agent based model 
and in turn make checkable predictions, and suggest further 
perturbations for refinement of a model.

Overall I'm just saying it is plenty hard to make predictions about 
relatively simple physical dynamical systems, even when its possible to 
poke them to see how they react.  Now let the particles have minds and 
layers of organizational insulation (receptionists, lawyers, etc.) and 
things get rather complicated when it comes to predicting things.

============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at http://www.friam.org

Reply via email to