|
Judith,
Ahhh, now I think
I see what you are saying. I myself would not call such a model also a
simulation, for a couple of reasons.
First of all, I
gotta be nitpicky. Rosen's definition of 'simulation' relies upon the concept of
'machine'. The toy airplane could not be a simulation
since it does not consist of a program running on a computer.
Instead, it could be called a simulacrum, if the
relation we are attempting to establish is between the overall behavior of the
toy plane and the corresponding behavior of the real plane, without regard to
establishing any synonymy of entailment relations between
them.
But if we are to
consider the toy plane a model (an analog, as opposed to a formal
model), then I would prefer to say not that the toy models some aspects of a
real plane and simulates others; but rather, that we choose observables
such that those observables are the ones we wish to model. If we put lead
weights at certain places on the toy plane in order to act as substitutes for
the weights of luggage, pilots, passengers, etc. on the real plane then we have
established a correspondence between the observables of aggregate
values, and locations, of these masses. I suppose we could equally say
that those lead weights are simulacra for the distinct masses in the real plane,
but the reason that they suffice is really because of the choice of observables
we make for the sake of that particular modeling relation. (This would be
typical in cases where we don't care about modelling any entailment
relations between, say, the seperate pieces of luggage - we don't need that
level of detail. Some other models, which must include, say, shifting
center of gravity due to loose luggage containers accelerated by sudden
dives due to windshear might require different observables and therefore would
not use aggregate values of total luggage weight in those MRs.) I
just think the terms simulacra/simulation for these kinds of observables clouds
the issue.
Regards,
Tim
|