I appreciate all of the comments I have seen. I certainly did not intend to ressurect the debate over whether accelerations cause forces. Let me attempt to rephrase my question:
If it is permissible in thermodynamics to say that gas molecules *could* do something (collect in one corner of a container) but are instead observed to do something else in every case (distribute themselves throughout the volume of a container) why can we not say that an object *could* have any acceleration but instead is always observed to have an acceleration which could be predicted by the equation a=Fnet/m?