- #36
Fra
- 4,207
- 629
I just skimmed the first page and I think the core issues here is the same as the various interpretations of probability.
As we know, entropy is a constructed (unique or not) measure of missing information, or alternatively a kind of measure of the probability of a macrostate given incomplete knowledge about the microstates, or a measure of the complexion number. If we take the subjective probability view here, then clearly also this entorpy measure is subjective (or observer relative), and thus there is a contradiction in imagining.
A frequent observation of entropy decreasing processes would be as to say that we frequently observer improbable events. Well, if we do repeatadly observer so called "improbable" events, then clearly the prior is revised and at some point it is not improbable anymore.
Anyway this all fits well, within a evolutionary view. If the observers are the ones that ENCODE the entropy measures, there should be an evolutionary selection in favour of "constructive codes". A encoding system that consistentely chooses to encode and retain only the most improbable events, would be extremely unfit/non-constructive and would not be very ubiquitous.
I think what he says that "physics cannot study those processes where entropy has decreased" is sensible if we interpret is as
"it's improbable that physical observers has evolved and persisted, that encodes it's own environment as improbable"
If anyone thinks this looks circular, then the circling, is the progress of evolution - there isn't necessarily a beginning and an end to everything. The physical basis of the measure "improbably past" as seen by any observer, simply doesn't exists in the actual past, since the observer itself is a result of evolution itself. So there is no need to pick infinitely unlikely initial conditions, since the complexion number itself, is reduced to triviality in the past, along with the observers.
/Fredrik
As we know, entropy is a constructed (unique or not) measure of missing information, or alternatively a kind of measure of the probability of a macrostate given incomplete knowledge about the microstates, or a measure of the complexion number. If we take the subjective probability view here, then clearly also this entorpy measure is subjective (or observer relative), and thus there is a contradiction in imagining.
A frequent observation of entropy decreasing processes would be as to say that we frequently observer improbable events. Well, if we do repeatadly observer so called "improbable" events, then clearly the prior is revised and at some point it is not improbable anymore.
Anyway this all fits well, within a evolutionary view. If the observers are the ones that ENCODE the entropy measures, there should be an evolutionary selection in favour of "constructive codes". A encoding system that consistentely chooses to encode and retain only the most improbable events, would be extremely unfit/non-constructive and would not be very ubiquitous.
I think what he says that "physics cannot study those processes where entropy has decreased" is sensible if we interpret is as
"it's improbable that physical observers has evolved and persisted, that encodes it's own environment as improbable"
If anyone thinks this looks circular, then the circling, is the progress of evolution - there isn't necessarily a beginning and an end to everything. The physical basis of the measure "improbably past" as seen by any observer, simply doesn't exists in the actual past, since the observer itself is a result of evolution itself. So there is no need to pick infinitely unlikely initial conditions, since the complexion number itself, is reduced to triviality in the past, along with the observers.
/Fredrik