“Atemporal” Extrapolations from Observations of the Past
A strategy Huw Price repeatedly uses in Time’s Arrow and Archimedes’ Point for showing that proposed explanations of time asymmetry in physical phenomena make implicit, unjustified time-asymmetric assumptions:
1. Starting from an atemporal standpoint, we have as much reason to believe that process X (the increase of entropy, for example) occurs in one direction of time as that it occurs in the opposite direction of time.
2. We know that process X did not occur in one direction of time (entropy did not increase towards the past, for example).
3. This throws doubt on 1. — on whether process X really does occur equally reliably in both directions of time. From a purely atemporal perspective, knowing that X did not occur in one direction of time, we should be compelled to think that it also does not occur in the other direction of time.
4. Therefore we have no right to think that X occurs in the other direction of time (that entropy increases towards the future, for example). Anyone who makes an argument to that effect is implicitly making a time-asymmetric assumption.
It is late at night and my brain is somewhat addled from struggling through the chapter on radiation (I never could get the hang of E&M), but I’m not convinced by this argument. I don’t think taking an atemporal perspective entails thinking that whatever we know to have occurred in the past has a significant chance of occurring in the future. Whatever we know has occurred in the past is derived from observation. It doesn’t necessarily tell us about laws or general trends leading from the past to the present, and hence, even taking an atemporal perspective, doesn’t necessarily tell us about laws or trends from the present to the future. So, sure, from an atemporal perspective, we are entitled to conclude that according to physical laws, the early universe should have a much higher entropy than the present universe. However, the fact (not based on laws but on observations) that the early universe had a low entropy cannot be used to tell us anything about what kind of entropic trend will continue into the future. In particular, it cannot be used to justify thinking, as Thomas Gold does, that the entropic trend we observed to continue into the past (decreasing entropy) will continue into the future. So ultimately we still have reason only, from considerations of physical laws, to think that entropy will increase into the future.
I may elaborate on this when my brain is less addled. But that’s it for now.
Yes, my brain is addled. Upon closer reading, all Price wants to conclude is that either we accept a symmetric Gold universe, or we accept that temporal asymmetry cannot be explained by physical laws (but is merely, for example, due to the very special initial conditions we observe). For if anyone comes up with atemporal physical laws that explain the initial state of the universe, those laws can be equally well applied to predict that the final state of the universe will be like the initial state of the universe. This recalls Price’s objection to dynamical explanations of entropy increase — they can never account for the asymmetry in our observations unless they themselves have asymmetric assumptions. So either we have a physical explanation that contradicts the facts (that indicate that our universe is temporally asymmetric) or we have no explanation at all.
I’d better go to sleep now before I type anything more idiotic.
March 30, 2007 at 3:17 pm
[...] and Archimedes’ Point) on some of the more philosophical questions of physics, with posts on Price’s objection to dynamical explanations of entropy increase (”they can never account for the asymmetry in our observations unless they themselves have [...]
April 1, 2007 at 9:44 am
[...] and Archimedes’ Point) on some of the more philosophical questions of physics, with posts on Price’s objection to dynamical explanations of entropy increase (”they can never account for the asymmetry in our observations unless they themselves have [...]