Monday, September 27, 2010 ... Deutsch/Español/Related posts from blogosphere

Bousso et al.: catastrophe imminent, time will end soon

Today, the arXiv harbors one of the most stupid hep-th preprints ever written:

Eternal inflation predicts that time will end
Raphael Bousso, Ben Freivogel, Stefan Leichenauer, and Vladimir Rosenhaus argue that the Universe has to abruptly die soon, in 5.3 billion years or so (it's their mean estimate). The whole world will probably disappear before the Sun runs out of its energy.

In the middle of the Sun's mature life, the atheist God of the Universe will suddenly say: "Sorry, dudes, time is over. I need to make at least one numerator in the papers by Bousso et al. finite, although not well-defined, so I have to kill you and everything else."

Their basic "reason" is simple. They think that the probability of an outcome is computed as a ratio of events in the multiverse where the outcome occurs divided by the number of all events:
p1 = N1 / N
And because both the numerator and the denominator are infinite numbers in an eternally inflating multiverse, they have to be made finite for various probabilities to be well-defined, they believe.

And the only way how they can imagine to make them finite is to euthanize the whole Universe, so they choose this as a "canonical" answer. ;-)
See also The Guth-Vanchurin paradox for a particular argument why the time should end, and its elementary flaws.
It's been amazing to watch these (and a couple of other) authors writing increasingly lousy papers and being transformed into downright crackpots during the recent 10 years or so. I am afraid that the transformation is already irreversible. These people are no longer capable to correctly answer even the most basic questions about the physical Universe.

Students, avoid these folks at any cost. They have become just a baggage of the Academia and a liability for the taxpayers. They managed to get stuck in the system as flukes - but the system is not yet quite broken and you will not be allowed to do the same thing.

The conclusion, which I have been presenting for years as a parody, incorrectly expecting that no recipient of a physics PhD degree could possibly make this argument seriously, clearly shows that the framework of their thinking about probabilities is fundamentally wrong.

Much like most other doomsday prophets and other kinds of crackpots, the authors are completely unable to understand the point and the arguments that they're assuming something fundamentally wrong in their very first axioms - and that's why everything they write down has to end up being a complete nonsense.

Instead, they obviously expect a long sequence of follow-up papers that will discuss whether the Universe will cease to exist in 5.31 or 5.32 billion years and worship the amazing prophets, Bousso et al., who have predicted this amazing doomsday. This is also clear from a staged dialogue included in the paper where a semi-dumb person asks various questions about the doomsday that the authors would clearly like to be discussed in other people's papers.

The paper also contains a lot of political kitsch - observers are systematically called "she" and subtle references to other kinds of doomsdays (the climate) are mentioned, too. As I mention below, this is probably what matters for the "uncriticizability of a paper" these days.

It probably makes no sense to explain these things again. The "typicality" crackpots won't listen, anyway.

There have been several related papers by Hartle and Srednicki that explain the right way to calculate the probabilities in a large Universe or multiverse; see also TRF (and the whole landscape category).

The main point is that a "xerographic distribution" - an assumption who we are and where we are in the Universe (or the likelihood that the answer is something or something else) - is a part of any scientific hypothesis about cosmology and this assumption must be tested (and may be ruled out) just like everything else we hypothesize. In particular, the assumption that "we must live at a random moment of time" in the history of a multiverse is just one possible xerographic distribution among others - and it is one that can be shown incorrect.

The Hartle and Srednicki papers - and others that make sense - are not even cited by Bousso et al. Bousso et al. don't want to listen. They continue to mindlessly push their rats through the mazes.

All of their key assumptions are wrong - and their paper is actually just another proof that they are wrong. It is not true that the probabilities of particular outcomes can be rewritten as quantities that are proportional to the volumes of space or volumes of spacetime. It is not true that for the concept of probability to be meaningful, the duration of the Universe has to be finite. Whether or not the eternal inflation has ever taken or is taking place, the assumptions about the probabilities requiring time to end is just wrong.

Eternal inflation does not imply anything of the sort.

After all, probabilities worked just fine before the 1920s when all the physicists thought that our Universe was eternal - more eternal than eternal inflation. It is not true that the probability is almost the same thing as the volume of space or spacetime. And even if you managed to interpret the probabilities in this naive "geometric" way, the resulting probabilities would be indefinite forms, "infinity/infinity", and the only correct regularization procedures would be those that would end up being compatible with the established laws of physics. In particular, any naive "geometric" cutoff of the Universe that someone needs to "visualize" the notion of probability is certainly unphysical.

The treatment by Bousso et al. is surely incompatible with all of physics.

First of all, it violates locality and causality. The probability that our Universe is going to do something in 5.3 billion years is given purely by the laws of physics (the quantum mechanical laws of the Standard Model pragmatically combined with general relativity, or the relevant portions of beyond-the-Standard-Model or string-theoretical physics, whenever hypothetically needed) and by the initial conditions - e.g. the present state of the Universe.

This setup is a complete toolkit and it contains the complete information that determines all the probabilities. So any additional framework you want to add to physics that implies different formulae for the probabilities - including the insane probability of a sudden death - is simply incompatible with physics.

Sometimes, you may be dissatisfied that some of our theories don't know everything - e.g. about the initial conditions. But be sure that if a framework knows "more than everything" and produces several different answers to the same question, it's inconsistent and therefore wrong which is much much worse than being "incomplete".

In other words, the setup of Bousso et al. is incompatible with causality. What will happen is yet to be seen. There can't possibly exist any influence of the hypothetical future (e.g. future collapses of the Universes) on the "current truth" i.e. on the right calculation of probabilities of events that will occur today.

Such a backward influence would be acausal: because the present era also and obviously influences the future, the influences in both directions would lead to closed time-like curves that imply all the well-known logical paradoxes. Our present phenomena - and the probabilities that some present events lead to particular outcomes - simply cannot depend on some questions from the future.

However, the Bousso et al. crackpot papers are full of these acausal mechanisms and calculations. For example, their calculations of the probability that it will be sunny today depends on some properties of the Universe (e.g. its volume) in the far future. They use insane formulae to calculate the probabilities of anything and everything which leads them to believe that the Universe has to die. Otherwise they couldn't compute the odds that it's gonna be sunny today. They can't compute them, anyway, but who cares. What matters is that the Universe has to die.

It's super-weird, isn't it?

Needless to say, Bousso et al. don't even bother to say anything about the "mechanisms" - i.e. the question how their claimed "end of time" will look like to a local observer and how it would fit into physics of particles and fields as we know them from the Standard Model and general relativity. For instance, what will be the shape of the hypothetical "cutoff of spacetime"? (Well, they suggest that only the Milky Way will suddenly disappear - an even better story.) They don't care because they don't give a damn about the actual physics. What they care about is their deluded doomsday philosophy that has nothing whatsoever to do physics.

Even if there existed some non-local or acausal effects in physics or in cosmology, they would have to be tiny not to disagree with the observations and they would have to be justified by some evidence that ultimately boils down to the empirical data. And all the empirical data we know can be phrased as properties of fields and particles that obey the well-known local and causal laws.

At this moment, there clearly exists no evidence supporting such bizarre effects that would directly or indirectly boil down to the empirical facts. The only evidence are arbitrary and incorrect philosophical assumptions about the probabilistic distributions in a multiverse - "probability should be proportional to the volume" - and there is an overwhelming evidence that these assumptions are wrong. Bousso et al. don't want to listen.

The paper also offers lots of arrogant statements such as that "the sudden end of time contradicts our intuition". It doesn't contradict just our intuition. It contradicts all of established science. All the particles and fields we have observed obey laws that guarantee that they will continue to do so. All of science is about having tested these laws and all of these laws imply that the processes basically have to keep on going: time cannot end.

The Universe may possibly suffer from instabilities - quantum tunneling into another place of the landscape, if you wish. But the only thing we can empirically know about the characteristic life expectancy of the Universes like ours is that it is almost certainly longer than 13.7 billion years. But just because this is what the simplest way to measure the life expectancy says, doesn't mean that the life expectancy has to be comparable to 13.7 billion years.

Don't worry: if you've read TRF just for 1 year, it doesn't mean that you will die in 1 year.

In fact, it's much more likely that the life expectancy is either infinite or at least vastly higher, something like 10^{100} years. If we ever learn the structure of the inflaton fields and their potentials, we will be able to calculate this number more accurately - or show why it's infinite (or Poincaré recurrence time which is basically infinite, around 10^{10^{100}} or googleplex of years). Although we don't know the exact answers, these questions may in principle be addressed scientifically. However, the scientific treatment has nothing to do with trying to make a manifestly wrong formulae or Ansätze for probability work.

I am not calling for any burning of the witches. But holy crap, if physicists don't lose all of their scientific credit by publishing this pure garbage and nothing else for years, can they lose their credibility at all? Does the institutionalized science have any checks and balances left? I think that all the people are being bullied into not criticizing the junk written by other people who are employees of the academic system, especially if the latter are politically correct activists. And be sure, some of the authors of this nonsense are at the top of it.

This is just bad. I urge all the sane people in Berkeley and other places to make it very clear to Bousso et al. - and to students and other colleagues - that they have gone completely crazy.

Add to Digg this Add to reddit

snail feedback (2) :

reader John Duffield said...

Good one Lub. Those guys just don't understand time. Clocks don't clock up the flow of time, they clock up motion. It doesn't matter whether it's a pendulum clock, a quartz clock, a light clock or an atomic clock. When you stop the clock you stop motion, not time. That goes back to Einstein's operational definition, heck no, it goes back at least as far as Aristotle. So what's going to happen, is the universe going to suddenly stop, just like that? Because anything can happen in the next five billion years? I've heard some garbage in my time, but this really takes the biscuit.

John Duffield

reader L. Edgar Otto said...


You have me thinking a little deeper or even more speculatively. I sort of see a link to your one on the probability paradox in a later post and the heart of all this human centered reasoning that really does not explain anything-especially the mechanisms as you point out.

But you read the articles on this before I did... so you may want to criticize or review my pesla.blogspot The Intelligible Multiverse

Effeciency is after all a facist word I hear but the lines in politics seems so unclear to me- it is as if we are entering a period where our era ends and the influence of the usa gets less to some vanishing point of global end of math. Perhaps it is the age of post modern philosophy of physics that we must get past as if Batilles symbol in the twenties near dadaism of a man without a head flirting with cannibalism but they seem to eat their own heads!

I guess in politics there are no necessary realities- how is it the women here have so many more degrees and still they make less pay? What does such blind correctness really tell us about our current university system?

The PeSla responding to a whole constellation of your new posts :-)

(function(i,s,o,g,r,a,m){i['GoogleAnalyticsObject']=r;i[r]=i[r]||function(){ (i[r].q=i[r].q||[]).push(arguments)},i[r].l=1*new Date();a=s.createElement(o), m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m) })(window,document,'script','//','ga'); ga('create', 'UA-1828728-1', 'auto'); ga('send', 'pageview');