Why is an improbable early universe unacceptable?

[student34]

We already observed it; it's too late to say it is too improbable. A particle from a star that exploded billions of years ago just went into my glass of water. But, if we ask the question before our sun blows up if a specific particle from it will land on the highest point of a planet 50 light years away, then we could say it wouldn't happen.

So I don't know why problems like the Boltzmann brain are actually problems. Any starting configuration, ordered or not, from a quantum fluctuation has extremely improbable odds, 1/(the number of possible beginnings). But it already happened, so the probability is 1.

What am I not understanding here?

 

[student34]

Crap, I just noticed the cosmology section.

sorry

 

[mathman]

What is your question? Distinguishing between low probability and impossible?

 

[student34]

From what I understand about the multiverse is that enough time has passed for a quantum fluctuation so rare that it had formed the beginning of this universe which resulted in what it has become today and will be in the future. Of all of the possible quantum fluctuations, ours is one of the most least likely - way too much order (low entropy). In other words, there are many other universes that should have arose instead of ours.

There are all kinds of explanations such as the anthropic principle that pretty much says that if we were any other universe, we wouldn't be here to ponder its rarity.

But my problem is why should it be strange in the first place. Nobody asked what the chances were of this universe before it happened, so why is it too improbable?

 

[Drakkith]

But my problem is why should it be strange in the first place. Nobody asked what the chances were of this universe before it happened, so why is it too improbable?

I'm sorry I'm having a hard time understanding your question. Why is what too improbable and what do you mean by that in this context?

Keep in mind that multiverse theories are extremely speculative, so I wouldn't put too much stock into any of them at this time.

 

[student34]

A major problem for cosmology is to understand how random quantum fluctuations began in such an ordered state, namely the Big Bang. But the probability of such a low entropy state is wayyyyy too small. We should be minds/brains that appeared just now, a second ago, hours ago etc where there is more entropy and less order. We should have scientists saying that the universe appeared at some state say 1 million years ago instead of one of the most improbable states of the Big Bang.

Here are two sides of the debate, string theorist and non-string theorist, respectively:




Both believe in inflation.

 

[Drakkith]

A major problem for cosmology is to understand how random quantum fluctuations began in such an ordered state, namely the Big Bang. But the probability of such a low entropy state is wayyyyy too small. We should be minds/brains that appeared just now, a second ago, hours ago etc where there is more entropy and less order. We should have scientists saying that the universe appeared at some state say 1 million years ago instead of one of the most improbable states of the Big Bang.

Seems to me that the universe may not exist the way it does just because of a quantum fluctuation. We have little to no idea of what physics was like during the very early universe, so we are certainly going to have issues explaining how the universe arose. It could be that the universe could only have ended up in a state similar to the one we are in. Or not.

 

[spacejunkie]

This issue is still confusing to the experts. I heard the same question being asked at the Time in Cosmology seminar at Perimeter.

 

[Chronos]

We have no way of knowing just how improbable our universe may be. Perhaps, given it's the only universe accessible for our observation and the only one of which we are aware, it is highly probable.

 

[student34]

This issue is still confusing to the experts. I heard the same question being asked at the Time in Cosmology seminar at Perimeter.

It's almost like they are saying that our universe is intrinsically special. But it's really just subjectively special to us because we

 

[student34]

My confusion is about why it's a coincidence in the first place. I don't understand why we should be surprised about something that we question the probability of after the fact. It's like if humans were floating in nothingness and hoping for this universe, then the only possible universe for them to live in popped out of nowhere. That would be a coincidence.

But it already happened. It happened, then we asked the question of its likeliness. I don't understand how this is a coincidence.

Lawrence points out in the lower video (at time 47:40), a possible solution is that we are naturally selected from cosmic evolution of other universes. A solution to what? Why would it require a multiverse? If this were the only universe possible, why should this universe be a coincidence?

 

[spacejunkie]

The idea is that we should be able to start a model from a position in phase space which is "typical" and get out a universe that looks like ours.

As it stands in GR the region of phase space that can produce one like ours is incredibly tiny compared to the total volume even if you can construct an appropriate measure on the phase space.

The multiverse idea just pushes the problem into a realm which can't be observed. Eternal inflation theories all suffer from the measure problem as do most, if not all, theories where space is non-compact.

 

[student34]

The idea is that we should be able to start a model from a position in phase space which is "typical" and get out a universe that looks like ours.

As it stands in GR the region of phase space that can produce one like ours is incredibly tiny compared to the total volume even if you can construct an appropriate measure on the phase space.

The multiverse idea just pushes the problem into a realm which can't be observed. Eternal inflation theories all suffer from the measure problem as do most, if not all, theories where space is non-compact.

The probability of the particles that make up my textbook was very low in coming together to form my textbook. It's not surprizing because I am thinking about it after it happened. I still don't see why we are treating our position in the phase space any differently.

 

[houlahound]

I think I see yr point; the probability for my ancestors to survive the plague, military conquests, two world wars, the depression, polio, crossing the road, both emigrating to a new country and meeting in a small rural town to produce my parents is mathematically so close to zero you can't tell the difference.

I do not understand this way of thinking about probability.

 

[Grinkle]

My confusion is about why it's a coincidence in the first place.

Its like spacejunkie says.

The idea is that we should be able to start a model from a position in phase space which is "typical" and get out a universe that looks like ours.

There is a dislike for theories that contain an element of extreme chance. The context for asking how likely something is comes from wanting to assess if a theory puts us at the receiving end of very special treatment. Its a measure of soundness of a theory, not some comment on whether our existence is likely. The likliehood that I am typing this post is 1, as you note.

So, if I theorize that process X caused our universe, and according to my models, process X nearly always causes things other than our universe, I might doubt that process X is indeed a good explanation for the existence of our universe.

Hope I read your question correctly.

 

[FactChecker]

The Boltzmann brain is not really a paradox. It is an example of the "self selecting sample".

Your OP example of a particle ending up in your glass of water is also not a paradox. If you pre-selected the particle and the glass before the experiment, it would be very unusual indeed if the pre-selected particle ends up in the pre-selected glass. On the other hand, if you pick a particle in the glass that "self-selected" itself to be in the glass, it is not a surprise that it ended up in the glass.

Similarly, if intelligent life asks why their universe is so lucky to support intelligent life, that is "self-selected".

 

[rootone]

I rarely bet on horses, but the first time I did that I got a winner at 40:1. picked it randomly just because it had intersting name.
However improbable an event may be, if the probabilty is greater then zero, then it can happen.

 

[houlahound]

Interesting debate;

https://www.theguardian.com/notesandqueries/query/0,5753,-6994,00.html

 

[Drakkith]

Interesting debate;

https://www.theguardian.com/notesandqueries/query/0,5753,-6994,00.html

I agree with this quote from one of the responders:

"I think there is a lot of unscientific nonsense floating about on this page."

 

[Chronos]

Assigning probabilities is a black art without prior experience to draw upon. After a billion coin flips the best you can deduce is whether or not the coin is fair. That in no way forbids a thousand heads, or tails in a row during that sequence.

 

[MathematicalPhysicist]

Assigning probabilities is a black art without prior experience to draw upon. After a billion coin flips the best you can deduce is whether or not the coin is fair. That in no way forbids a thousand heads, or tails in a row during that sequence.

You can never really know if a coin is fair or not by only observing the outcome of tosses.

 

[windy miller]

The idea is that we should be able to start a model from a position in phase space which is "typical" and get out a universe that looks like ours.

As it stands in GR the region of phase space that can produce one like ours is incredibly tiny compared to the total volume even if you can construct an appropriate measure on the phase space.

The multiverse idea just pushes the problem into a realm which can't be observed. Eternal inflation theories all suffer from the measure problem as do most, if not all, theories where space is non-compact.

Surely any model of the universe that have an infinite amount of time have a measure problem. Why shouldn't that also be true of standard big bang cosmology without inflation. After all , current data implies our universe will expand forever and so there is a infinite amount of time in its future, so this picture also has a measure problem.

 

[houlahound]

Not sure if a future infinity is an actual thing that you can base calculations on.

 

[windy miller]

Consider the Boltzman brain problem, if you wait long enough in a accelerated expanding universe you might expect BB's to form. Do you need eternal inflation for this problem ? No you don't. Just wait long enough, so now you have the problem of why aren't you a BB? This problem is just as much a problem for single universe models as it for eternal inflation, so single universe models have a measure problem too.

 

[houlahound]

Please define what a measure problem is.

 

[FactChecker]

Assigning probabilities is a black art without prior experience to draw upon. After a billion coin flips the best you can deduce is whether or not the coin is fair. That in no way forbids a thousand heads, or tails in a row during that sequence.

In cases where there is a theoretical basis for determining probabilities, prior experience is not needed. In the case of coin tosses, there are many statistical tests that can be done to see if it is believable that the coin is fair. A thousand consecutive heads in a billion is probably a good indicator that the coin tosses are not fair. For one thing, the lag 1 autocorrelation would be high. The probability of such a high autocorrelation from a fair coin would be very low. (I have not calculated it.)

 

[Chronos]

Yes, a run of 1000 is highly improbable, but, a run of 30 would be entirely unremarkable. Even several such runs would not be particularly significant.

 

[FactChecker]

Yes, a run of 1000 is highly improbable, but, a run of 30 would be entirely unremarkable. Even several such runs would not be particularly significant.

Right. My point is that there can be statistical tests based on the theory of the proposed distribution that can be applied. It does not necessarily depend on prior experience or data.

 

[Chronos]

What distribution model applies to a sample size of one? We have zero information, theoretical or otherwise, governing the variance, or if any even exists for the fundamentals of our universe. Even a crooked coin looks pretty random more often than not.

 

[FactChecker]

What distribution model applies to a sample size of one? We have zero information, theoretical or otherwise, governing the variance, or if any even exists for the fundamentals of our universe. Even a crooked coin looks pretty random more often than not.

Almost all distribution models give the probability of a single sample of the random variable. Suppose a theoretical distribution has a standard normal distribution with mean=0 and std dev=1. If I get a single sample value of 1000, then I can say with great confidence that it does not come from that theoretical distribution.
But I don't want to hijack the thread. I just saw a statement about probability that I couldn't resist commenting on.

 

[spacejunkie]

Please define what a measure problem is.

A measure problem is either when different measures applied to the same space in a theory produce different predictions or there is no way to define a measure on a space at all.

Measures in set theory and topology are ways to assign numbers to subsets. They generalise lengths, areas and volumes.

https://en.wikipedia.org/wiki/Measure_(mathematics)

Measures also apply to probability spaces so come into topics like entropy and quantum theory.

Measure problems are endemic in cosmological theories and many different regularisation procedures have been created to overcome them but none are generally accepted. In eternal inflation theories the problem is how to apply probability measures to infinite sets of causally disconnected universes. In canonical quantum gravity the Wheeler-deWitt equation is regularisation dependent.

https://en.wikipedia.org/wiki/Measure_problem_(cosmology)

Please define what a measure problem is.