Modelling

[Marg0]

What are the rules for creating models in science and comparing them to the outcome of experiments? For example I can create a computer model of how I would expect a certain type of thought process to occur. How can I be sure my reasoning does not just become isolated or circular?

[ucbprja]

The answer to your first question is that there aren't any rules (aside from ethical ones!) - only peer review! Of course if you model a particular process then presumably you want to you use your model to test a hypothesis, i.e. ask an interesting scientific question. To validate your answer you usually need to measure some outputs of your model and compare to experiment, although you could have an a model constructed on the basis of a priori assumptions, like special relativity.

Your second question I guess is philosophical one? I can't really seeing it being a practical problem, because if you have a state of the art model of the brain then it's still a crude approximation, and hence there is no way you'd be concerned with your reasoning being circular. If in the future you had a perfect model then it would be a copy of your brain, and that would resolve the circular nature of the argument.