The book Thermodynamics and Statistical Physics has an interesting section on "Entropy and Universal Accidents" (pp. 120-122). Some excerpts:
Entropy increases until it reaches a maximum value. The irreversible processes responsible for the increase of entropy are the same mechanisms which drive the system toward equilibrium...
However, the myriad of irreversible processes which perform so ably in establishing equilibrium in subcosmic chunks of the universe seem to lose their punch when applied to that part of the universe visible to us...
One suggested resolution of this paradox is known as the fluctuation hypothesis...by Ludwig Boltzmann...
A fluctuation in a thermodynamic system is the result of an accident on the atomic level involving many atoms...Such fluctuations occur by chance--they are accidents of nature...
Boltzmann's fluctuation hypothesis suggests that the universe is in equilibrium but that the portion which we observe is part of a gigantic fluctuation--the granddaddy of all accidents...However, the argument is unsound. It is enormously more likely that such a fluctuation would occur over a small volume, say the size of our solar system, and leave the rest of the immediate universe in equilibrium...
How then can we explain the fact that we live in a restless universe? One possible answer is suggested by the theory of relativity...[which] teaches us that the geometry of space is determined by the masses within the space. As the distribution of mass in our universe varies so do the geometrical properties of space...Such a system would not approach equilibrium because the external conditions are changing.
A couple thoughts:
1) In purely physical terms, it would seem that "localized" non-entropic systems are accidental (insofar as they go contrary to the "direction" of irreversible processes tending toward equilibrium). And this would hold true as well on the atomic scale, e.g.: the formation of molecules usually requires overcoming the activation energy of the chemical bonding. Without some external influence (H2 + O + spark ---> H2O; or, the presence of enzymes or catalysts), such compounds would ordinarily not be formed. But quantum mechanics assures us that, however small the probability may be, it is possible for such compounds to form spontaneously, solely by chance (as indeed is observed, though the formation rates may be very low). It seems to me that both the cosmological and atomic processes involved here are completely random.
2) Suppose that it is the relativistic variations of local space that "drive" these observable fluctuations. Then that too is a totally random process, unless we assume that someone is choregraphing the distribution of matter in space so as to produce "favorable" environments.
If this is so, then from what I've read (yes, including that boogered but not completely erroneous Creation book) it would seem virtually impossible that even one strand of DNA would ever form as a result of such purely random processes. Thus my feeling that there is some as yet unidentified force that made this all happen.
Craig