by smiley » Fri 19 Nov 2004, 11:14:07
$this->bbcode_second_pass_quote('', 'A')pplying this knowledge to nature, physicists found that the total entropy change (change in S) always increases for every naturally occurring event (within an isolated system) that could be then observed. Thus, they theorized, disorder must be continually augmenting evenly throughout the universe.
When they applied this to nature they found that everywhere the order increases and in order to justify that observation they assumed that somewhere else in the universe the disorder must increase in order to let the total entropy increase. Only so far they haven't been able to point out where exactly that might be happening.
But order and entropy don't bite. Look at the original Boltzmann equation,
S = k ln W (whereby W is the number of complexions)
Imagine that you have a deck of cards and you draw 4 cards from this deck. Then W should be interpreted as the total number of possible combinations. Most of these combinations are worthless, only a few have value. Each of the combinations has an equal likelihood of occurring
Because there are more "worthless" combinations of cards than "valuable", the probability that you draw a bad hand is larger than that of a valuable hand. This example illustrates the probabilistic nature of the disorder argument. Separate ordered and disordered states have a equal likelihood of occurring. Simply because there are more disordered combinations possible, disorder is more likely to occur. Moreover it shows that disorder and entropy are really separate issues as the entropy is solely defined by the number of elementary complexions, not by whether these complexions are ordered or unordered in nature.
The above considerations are made on the premises that each complexion has an equal preference. However this only holds when the subjects do not interact, like in a perfect gas. The Clausius assumption is thus a special case in which interactions can be ignored. When you have interaction some states will be favored over others, and these are usually ordered states. The W that we use now does not reflect that preference.
This argument is a very old one as it was first expressed by Albert Einstein. Unfortunately his argument was hijacked by the creationist bunch to justify their view of the creation. But the original argument does still hold. When you take in account the orbital theory and the hybridization of the atoms you can understand why ordered states are favored in cases of strong interaction.
Experimentally, chaotic systems are very hard to maintain. A small push will bring them to order. Nanoseconds after the big bang the universe was in ultimate disorder, a hot sphere of expanding gases. Yet a miniscule imperfection made the the universe evolve in the rather ordered state we know now. Chaotic systems only need a very small seed to resolve into order. The opposite is not true.
Ordering phenomena are usually explained in terms of entropy decreasing locally while the entropy of the universe as a whole increases. This is a very poor interpretation since it is possible to increase the order in an isolated system. If you put some liquid in a perfectly thermodynamically sealed box it can crystallize. When you assume that both entropy and order can increase at the same time the interpretation of the events around us becomes a whole lot easier.
The problem with debating these things has grown historically. Thermodynamics have always created a lot of flak from the creationists, who have seen thermodynamics as a direct attack on their image of the universe. They have tried to use whatever they could find to discredit thermodynamics. This has caused the people who believe in thermodynamics to staunchly defend their ideas. Unfortunately this has also caused them to become as blind as the creationists.
It seems that this debate has moved to the background as more and more people seem to be convinced of the need to refine our idea of entropy. If you interested in the subject I would recommend you to read more of the work of Tsallis. The work still needs to be refined, but it has shown its usefulness in many cases where the Boltzmann equation fails. I think it is a very big step forward.
http://www.santafe.edu/research/publica ... sallis.php