Inflation as a solution to the Early Universe Enrtopy problem

[skydivephil]

I really think this paper
http://arxiv.org/abs/1212.1087
which Chronos mentioned in another thread deserves its own thread. Has anyone had a chance to look at it? what do you think? This has been a lot of noise about this issue raised by people like Roger Penrose and Sean Caroll . They have come up with the some pretty , shall we say, creative ideas such as CCC and the Caroll/Chen model to deal with this issue. But if this paper is right they are not required.

 

[yenchin]

I really think this paper
http://arxiv.org/abs/1212.1087
which Chronos mentioned in another thread deserves its own thread. Has anyone had a chance to look at it? what do you think? This has been a lot of noise about this issue raised by people like Roger Penrose and Sean Caroll . They have come up with the some pretty , shall we say, creative ideas such as CCC and the Caroll/Chen model to deal with this issue. But if this paper is right they are not required.

It certainly is an interesting idea, i.e. to appropriately take account into quantum [entanglement] entropy. But personally I don't believe the problem is quite solved. Recall that the problem we had was how to get inflation started in the first place. It may be true that indeed the probability of inflation does somewhat increases if we were to take into account various quantum effect. However part of the argument in the paper relies on the fact that inflation is always accompanied by particle horizon [and so there is entanglement entropy of stuffs beyond and in the horizon]. However, that means inflation has already started! One can of course think of some other way out of this and show that inflation can indeed happen once all quantum effects are properly taken into account; so I do feel that it is a fresh idea in the field.

 

[Chalnoth]

I really think this paper
http://arxiv.org/abs/1212.1087
which Chronos mentioned in another thread deserves its own thread. Has anyone had a chance to look at it? what do you think? This has been a lot of noise about this issue raised by people like Roger Penrose and Sean Caroll . They have come up with the some pretty , shall we say, creative ideas such as CCC and the Caroll/Chen model to deal with this issue. But if this paper is right they are not required.

At first glance, it sounds a bit unlikely. The claim is that inflation itself lowers entropy over time, which is another way of stating that they're using an invalid measure of entropy.

That said, I don't think you should lump in CCC with the Caroll/Chen model. The Caroll/Chen model is fairly reasonable. The CCC model is way, way, way out there.

 

[Chronos]

They are using von Neumann entropy as a measure of entropy, which is the quantum version of entropy. This appears entirely reasonable to me. The Carroll/Chen model, http://arxiv.org/abs/hep-th/0410270, invokes eternal inflation, which has its own issues.

 

[skydivephil]

It certainly is an interesting idea, i.e. to appropriately take account into quantum [entanglement] entropy. But personally I don't believe the problem is quite solved. Recall that the problem we had was how to get inflation started in the first place. It may be true that indeed the probability of inflation does somewhat increases if we were to take into account various quantum effect. However part of the argument in the paper relies on the fact that inflation is always accompanied by particle horizon [and so there is entanglement entropy of stuffs beyond and in the horizon]. However, that means inflation has already started! One can of course think of some other way out of this and show that inflation can indeed happen once all quantum effects are properly taken into account; so I do feel that it is a fresh idea in the field.

There have been a number of papers which suggest the probability of inflation is 1 in LQC, see here:
http://arxiv.org/abs/0912.4093
http://arxiv.org/abs/1011.4249

So it seems to be possible both of these results are valid, if so might a solution then have been found? BTw on the CCC model and Caroll/Chen models, I was simply trying to ascertain if this paper (perhaps combined with the results in the 2 paper I mention above) removes the need for such models. How reasonable or unreasonable they was not my issue here.

 

[yenchin]

There have been a number of papers which suggest the probability of inflation is 1 in LQC, see here:
http://arxiv.org/abs/0912.4093
http://arxiv.org/abs/1011.4249

So it seems to be possible both of these results are valid, if so might a solution then have been found? BTw on the CCC model and Caroll/Chen models, I was simply trying to ascertain if this paper (perhaps combined with the results in the 2 paper I mention above) removes the need for such models. How reasonable or unreasonable they was not my issue here.

Yes, I am aware that LQC does that, but what I don't understand is: if we accept LQC explanation, then isn't that all there is to it? As long as we can get inflation to start, then what is the problem? [Since the original problem is that inflation itself does not explain the arrow of time because inflation needs even lower entropy to get started, i.e. we need to either explain how the inflation starts despite generic initial condition (which LQC probably does), or explain why the initial state is so different (such as http://arxiv.org/abs/0711.1656v2)] .

The whole point of the paper, as I understand it, is that they are trying to explain how inflation can get started entirely within the quantum entanglement framework.

 

[Chalnoth]

They are using von Neumann entropy as a measure of entropy, which is the quantum version of entropy.

...of a generalized Chaplygin gas, which is a highly hypothetical material. And stating that they're using von Neumann entropy is a relatively meaningless statement. Von Neumann entropy is merely a counting of the microstates: what matters is what microstates they are counting. An invalid measure of entropy is one in which they count the microstates in an invalid way.

 

[yenchin]

...of a generalized Chaplygin gas, which is a highly hypothetical material. And stating that they're using von Neumann entropy is a relatively meaningless statement. Von Neumann entropy is merely a counting of the microstates: what matters is what microstates they are counting. An invalid measure of entropy is one in which they count the microstates in an invalid way.

What is your criteria of validness for measure?

 

[Chalnoth]

What is your criteria of validness for measure?

Entropy tends to increase over time, following the second law of thermodynamics in the statistical sense as derived from statistical mechanics.

 

[yenchin]

Entropy tends to increase over time, following the second law of thermodynamics in the statistical sense as derived from statistical mechanics.

I do agree with you. However in the paper the authors mentioned "We notice that in the literature there are examples of decreasing entanglement entropy [15][16] due to the situation where the system is out of equilibrium or simply from long-range quantum correlation". So I am somewhat confused.

 

[yenchin]

Yes, I am aware that LQC does that, but what I don't understand is: if we accept LQC explanation, then isn't that all there is to it? As long as we can get inflation to start, then what is the problem? [Since the original problem is that inflation itself does not explain the arrow of time because inflation needs even lower entropy to get started, i.e. we need to either explain how the inflation starts despite generic initial condition (which LQC probably does), or explain why the initial state is so different (such as http://arxiv.org/abs/0711.1656v2)] .

The whole point of the paper, as I understand it, is that they are trying to explain how inflation can get started entirely within the quantum entanglement framework.

I now understand the paper better after I discussed with the authors [well, actually... one of them is my supervisor and I also know the other two] The point of the paper is to point out that if we take into account entanglement entropy, then according to their way to counting, there is a larger probability for inflation to happen [but if you believe the LQC people, then the problem is already solved]. That is, they don't really aim to explain *how* inflation gets started, merely that there is a larger probability that it *can* happen than traditionally believed by, e.g. Penrose.

 

[skydivephil]

As I understand there are two issues that we would like answers to (well obviously more than 2 but 2 related to this thread):
1 how did inflation get started and how likely was that state?
2 why was the entropy of the early universe so low?
From my reading the paper attempts to address question 2 and a paper like Ashtekar/Sloan addresses question 1. So assuming 1 is solved I don't think that means that 2 is solved or am I wrong?

 

[yenchin]

To be frank, I am not sure if explaining inflation is sufficient, but one way to think about the issue is as what McInnes mentioned in http://arxiv.org/abs/0711.1656v2.

Inflation gives us an enormous simplification of the problem of explaining the Arrow. As Huw Price has emphasised [11], we are by no means entitled a priori to assume that the low initial entropy of our Universe was “stored” in a simple way or confined to a single form of matter. Inflation performs precisely this service: it implies that the entire, potentially vastly complicated problem of accounting for the entropy of the early Universe reduces to the single problem of accounting for the initial state of the inflaton. In the words of Lisa Dyson, Matthew Kleban, and Leonard Susskind [12]: "Some unknown agent initially started the inflaton high up on its potential, and the rest is history." In fact, at stake here is the reason why there was any history: if we can understand the "unknown agent", then, thanks to Inflation, we will have a theory of the "passing" of time.