From PhilPapers forum Philosophy of Physical Science:

2011-12-16
Are the physically possible worlds the same as the logically possible worlds?
Ramon, your position is more or less the conclusion I came to in the form that limitless logical possibilities cannot be imputed to an analogous physical system, for a physical system is subject to Second Law.

In my own field, disregard of this can be crippling. A systems approach is almost always explicit or implicit in the work of social scientists in that they rely on a causal relation of entities or the interaction of reified factors or parts that somehow function to support an emergent whole. This forces them to disregard your point in order that they might salvage human freedom and moral responsibility from the fact that all systems (i.e., a complexity framed as an entity , hypostatized as an isolate, or having an initial state defined as entirely observable) are necessarily mechanical and determinant. They are at a loss how to reconcile this with "free will".

This comes up in a remarkable variety ways which attempt to get around the fact that all systems in principle are mechanistic ("open system" is a contradiction of terms). For example, I frequently encounter the chaotic system argument that recursion makes a system is indeterminant. The semantic argument is made that the limitless possibilities for words and their different combinations in different contexts is the basis of human creativity. Also, since system complexity yields unpredictable outcomes because of our epistemological limitations, it is taken to imply that the physical system is itself ontologically indeterminant.  

Haines Brown