1

Statistical mechanics says that a system will evolve to a state of higher entropy (i.e. states with higher number of microstates) simply because it is overwhelmingly more probable than evolving to a state of lower entropy. Does this simply that the evolution of a system is inherently probabilistic?

Say we have a container of gas. According to kinetic gas theory, we can model the gas as a collection of particles that obey the law of Newtonian mechanics. But since Newtonian mechanics is deterministic (i.e. if we know the initial condition, we can know any future state with 100% certainty), the container of gas as a whole should behave deterministically. If this is the case, why do we assign probabilities to different states of a system? Is it because we don't exactly know the initial condition of a given system that we resort to probabilities?

1 Answers1

2

No. Statistical mechanics does not imply a non-deterministic universe.

When you ignore some information about a deterministic system you can still use probabilities. Consider a machine that flips a coin to the opposite side every 5 seconds. If we don't know how long since the machine was started, we can say there is a 50% chance of the coin being heads when we look.

Consider a machine that starts with 8x8 coins all heads, and every 5 seconds it uses a pseudorandom number generator to select a coin and flip it over. After the machine has been running for a long time we can estimate that about 50% of the coins will be heads. In principle we could calculate it using knowledge of the pseudorandom number generator and the amount of time since the machine started running. In reality we probably don't really care about calculating it like that, and we just use a probability distribution to describe the likely number of heads showing at any one time.

user253751
  • 1,496