Hello,
I'm following the discussion a while:
>The contributions of the past few days seem to point up the
>concatenation of the concepts of prediction, constraint and the
>relationships between individual behaviour and social phenomena.
...but I'm becoming a little bit lost in it. I have following questions:
1. What do you, folks, mean by prediction?
I pose this question because it seems to me that this concept was used
here much in the sense that we in our heads, so as to say mentally,
according to our empirical knowledge try to know in advance (predict)
all the possible future actions agents in the model may take and
consequently all the possible states of the population of agents and its
environment. (do I misinterprete?). Instead I think that the prediction
is (one of) the reason d'etre of simulation models and is the result of
'running an experiment'. To verify the results statictical data can be
used. So I think the question of more or less constrained models of
agents and the environment they are situated in is important only from
the point of view of a.) the computational capacity of hardware being
used and b.) relevace of constraints for modeling some real phenomena.
I think that there are two major ways to get useful models of social
phenomena:
- to model individual empirical models of agents' behaviour and their
environment and test what will emerge from their interactions (both
among agents and among agents and the environment). This is the
bottom-up approach
- to model the agents and the environment so as to get the 'required'
results.
from the statements above emerges :-) my second question:
2. What are the implications of
>This in turn means that it is impossible to predict anything without
>(a) a theory of social entities, their functions, and their effects on
>agents' behaviors. But on the other hand, it is pointless to try to
predict
>these effects without a
>(b) a theory of how agents
>- form the social structures they are involved in. This cannot always
be
r>educed to a behavioral effect (like queuing models). Sometimes,
patterns
>of social relationships, networks, etc. _emerge_ from agents' cognitive
>properties, be the agents aware of this or not. Different agents with
>different mental features (e.g., goals and abilities) will form
different
>patterns of interdependence. If one does not model agents' mental
>diversity, one does not fully understand a fundamental source of social
>embeddedness.
>- Adapt/contribute to make social institutions, etc. effective. Whether
>more or less constrained, deliberative agents have the capacity to form
and
>break their commitments, to accept and drop social collective
intentions,
>to form and revise their (social) beliefs, and have the capacity to
decide
>whether and to what extent to support a given norm, institution, etc..
Now,
>these decisions are _not_ unaccountable, nor unpredictable, nor purely
>ideosyncratic. Of course, to account for them makes the theory of
social
>forces more complex and less straightforward: but this only means that
we
>must refine our theoretical instruments! To account for the _link_
between
>the two levels of complexity (mental and social) is always necessary
in
>order to fully understand either level.
for computer models of social phenomena?
I hope this was not astray and thanx for time spent
Ales Kubik
************************************
Ales Kubik
Department of Applied Informatics
University of Economics
Bratislava
Slovakia
1001110> [log in to unmask]
to hear my voice: 00421 (0)7 67295 865
************************************
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|