I’ve been critical of multiverse pseudo-science because it doesn’t make any testable predictions, but it seems that tonight there really is one. According to this new preprint, multiverse arguments guarantee that time will end, with the expected amount of time left before the end about 5 billion years and
There is a 50% chance that time will end within the next 3.3 billion years.
The argument seems to be that multiverse arguments require introducing an artificial cut-off to get finite numbers, so the cut-off must be there and we’re going to hit it relatively soon on cosmological time scales. The age of the universe is about 13.75 billion years, but we’re getting near the end, already entering late middle-age to senior-citizen time-frame. One interpretation given of this result is that:
we are being simulated by an advanced civilization with a large but finite amount of resources, and at some point the simulation will stop.
It turns out that you don’t even need the whole apparatus of eternal inflation to see that time is going to end. All you need to do is to think about sleeping and waking up, which, according to the paper, leads to the “Guth-Vanchurin” paradox:
Suppose that before you go to sleep someone flips a fair coin and, depending on the result, sets an alarm clock to awaken you after either a short time or a long time. Local physics dictates that there is a 50% probability to sleep for a short time since the coin is fair. Now suppose you have just woken up and have no information about how long you slept. It is natural to consider yourself a typical person waking up. But if we look at everyone who wakes up before the cutoff, we find that there are far more people who wake up after a short nap than a long one. Therefore, upon waking, it seems that there is no longer a 50% probability to have slept for a short time.
How can the probabilities have changed? If you accept that the end of time is a real event that could happen to you, the change in odds is not surprising: although the coin is fair, some people who are put to sleep for a long time never wake up because they run into the end of time first. So upon waking up and discovering that the world has not ended, it is more likely that you have slept for a short time. You have obtained additional information upon waking – the information that time has not stopped – and that changes the probabilities.
However, if you refuse to believe that time can end, there is a contradiction. The odds cannot change unless you obtain additional information. But if all sleepers wake, then the fact that you woke up does not supply you with new information.
Update: Lubos doesn’t think much of the paper:
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.
Update: In other pseudo-science news, the latest Scientific American features a piece by Hawking and Mlodinow based on their recent book.
Update: Not only does New Scientist think this nonsense deserves to be covered in a lead article, they also have an editorial urging us not to “roll your eyes” about this.
The cutoff is a computational device – I don’t see any physical reason for a discontinuity between the region within the cutoff and the region outside the cutoff (figure 1). All that appears to be happening is that no matter how I construct an ensemble of observers in the eternally inflating universe using a geometrical cutoff, a non-zero fraction of observers will hit the boundary. I don’t see how that is interpreted as time coming to an end. In particular, I can have a universe in which every observer sees 1 o’clock and 2 o’clock, but there is no way for me to select a geometrical cutoff that contains only observers that see both events, i.e., a vanishingly small fraction of observers who see only one o’clock. If I believe the paper, then my cutoff actually truncates the multiverse, which is a god-like power to ascribe to physicists.
Um, and the arxiv bans papers on Koide formulae for neutrino masses? What universe am I in?
Ok, I’m going to try very hard to be polite, and take this statement seriously: “There is a 50% chance that time will end within the next 3.3 billion years.”
Why is this prediction testible? Isn’t there a 100% chance that anyone who takes this bet will not live long enough to find out if the prediction is accurate thus negating the meaningful concept of making a prediction?
If we are a simulation, we have no way of knowing what kind of laws of physics rule the universe of the beings that simulate us. It might very well be that in their universe conservation of energy does not hold, so they have in fact infinite resources for running the simulation. Perhaps the whole simulation is done to figure out how an intelligent species would adapt to a universe where resources are limited.
When these people teach, what do the students gain? Does anyone have first-hand information, I wonder.
This is an easy paper to mock, and I guess deservedly so.
Section 5.1 contains a clue to the paper they should have written. It lists assumptions, one of which must be incorrect, for the conclusion that “time ends” to be false. The second one is “Probabilities in an infinite universe are defined by a geometric cutoff”.
So the paper they should have written would have been “Probabilities in eternal inflation can’t be defined by a geometric cutoff”, and this whole business about “time ending” would have been a one-sentence reductio ad absurdum of a particular line of argument.
Amazing how far one can drift away from rationality when one is surrounded by equally deluded peers.
The sloppiness of thinking in the paper is really remarkable, here is an example from the second paragraph:
“If it does occur in Nature, eternal inflation has profound implications. Any type of event that has nonzero probability will happen infinitely many times, usually in widely separated regions that remain forever outside of causal contact. This undermines the basis for probabilistic predictions of local experiments. If infinitely many observers throughout the universe win the lottery, on what grounds can one still claim that winning the lottery is unlikely? To be sure, there are also infinitely many observers who do not win, but in what sense are there more of them? In local experiments such as playing the lottery, we have clear rules for making predictions and testing theories. But if the universe is eternally inflating, we no longer know why these rules work.”
Is this for real? The rules work because the lottery is designed that way! For every winner there are thousands of non-winners, the infinite number of lotteries doesn’t change a damn thing.
It’s like claiming that if eternal inflation is true we no longer know why the probability of rolling 6 on 6 sided dice is 1/6!
And even if we ignore the atrocious lottery example and talk about probability in general the paragraph still makes no sense – we have an empirical way of defining probabilities and it works very well and is completely immune to eternal inflation metaphysics.
As for the main conclusion the fact that eternal inflation requires an ad hoc regulator to make any predictions means that it is a useless pile of philosophizing not that the time will end.
The authors run a serious risk of not being taken seriously. Is the universe inflating infinitely or merely the egos of those involved?
In fact maybe I should refrain from commenting as I admit to not reading more than the first paragraphs; I got stopped early on by the quotation cited by Nex, which t is so clearly complete BS. On the other hand, in infinitely many universes I will apparently “lose the lotto” and continue reading. In infinitely many of these same universes these authors will win jobs from places like Berkeley, Lawrence Labs, and be paid actual salaries…….
I close with a question, which should be testable: if physics articles on the arxiv keep inflating, won’t the average information content approach zero exponentially fast???
So I wake up, and look at the others who’ve just woken up and those who are still sleeping, and see that it’s 50/50. Whether or not “time will end.” Don’t see what’s paradoxical about that.
I agree with Lubos. What garbage! The gist of their argument seems to be that because something (the end of the time) CAN happen in a large enough sample it necessarily WILL. MIT and Berkeley, eh? It may not mean the end of time, but it does seem to indicate the end of any standards in theoretical physics research.
I see this kind of “reasoning” all over the place from multiverse advocates and it’s not just bad science, it’s bad logic. Even my five year old son understands that when you get one random pick, “you get what you get and you don’t get upset”.
The glaring flaw is the assertion that “it’s natural to consider yourself a typical person”. THIS IS COMPLETE NONSENSE, and this cannot be repeated often enough and loud enough. Basically, we are taking a sample of size one from an allegedly very large population, and I know of no principle in probability that justifies one in drawing *any* conclusion whatsoever from this. (And no, the “Copernican principle” will not do).
In particular you cannot, as multiversians are fond of doing, reason about the population you were drawn from based on one outcome. Consider this analogy: you have a large sack of balls, some black, some white. You draw out one ball and it’s white. What can you conclude about the proportion of black and white balls? Answer: Absolutely nothing. (Well, OK, you can conclude that there was at least one white ball in there…). You can’t even conclude that there are many universes, let alone that ours is typical.
OK, I’m sorry, it used to be fun for mathematicians like myself to make fun of physicists, for example I was just at a conference celebrating Scott Wolpert’s 60th birthday, and Peter Sarnak made a crack “he’s a physicist, so he doesn’t have to prove anything” about someone applying percolation theory to questions on nodal domains of cusp forms, and it got a big laugh. But stuff like this just ruins it all, I mean how can you make fun of people who take stuff like this seriously, at this point I’m just embarrassed for you all 🙂
Oh no, they didn’t repeat the old nonsense about “in an infinite multiverse, everything must happen infinitely many times”, did they? I dispatched of that chestnut in the *letters column of New Scientist*, for crying out loud. Could some friendly, patient mathematician explain to these physicists that infinities come in more than one size?
Not to mention the fact that if eternal inflation is driven by bubble universes born from existing ones, it’s perfectly reasonable that the bubbles have physical laws not far distant from the parent, and that however often the bubbles are created, they never explore more than a tiny fraction of possible universes. It’s even possible that they cycle through a finite set of configurations. All of this is at least as likely as the “everything must happen” nonsense.
The logic in this paper is a disaster. This “Guth-Vanchurin” paradox excerpt given here is just begging the question, and so is the assumption that the theory must result in “finite numbers”. The logic is: I want finite results, so I will assume that finite results are correct, and then write a paper concluding time is finite.
I’m a liberal arts major. Come on, physicists.
The ‘Guth-Vanchurin paradox’ as written up by these people seems like nonsense but there is a very similar real effect in statistical physics based on Levy statistics.
In fact, in the last year I read of an experimental paper in PRL that observes the effect in some quantum dots.
Here is the essence: Imagine a particle (or quantum dot) with two states – glowing and not-glowing. The particle makes stochastic transitions between these two states and the amount of time spent in each of these states comes from two distributions G(\tau) and NG(\tau).
Let, say, G(\tau) have a long tail (like a Levy distribution). That is, there is significant probability of spending very large times in the glowing state. Let’s say NG(\tau) does NOT have such a tail.
Then, if you take a population of such dots and observe the total brightness, you’ll find that the total intensity actually increases with time even though the properties of each dot are time-invariant. This is because the longer you measure, the more of the tail of G(\tau) you start sampling…
The essence is that when you have a distribution of times like G(\tau), the length of the experiment you sets a growing cutoff on the distribution. The paper in PRL measures this effect in a real system.
My philosophical resolution of the paradox is that it is not possible to verify that a particle has the distribution G(\tau) in the first place without waiting a long time.. or something of that sort — what does it really mean to have a long tailed distribution of times and how would you verify it?
Do you also object to predictions that our Sun will explode into a red giant in a similar time-frame? It is possible for a testable theory to make long-term predictions, so it is not enough to just argue that people will not live that long.
“How can the probabilities have changed?”
Because you changed the sample population and in doing so biased it arrrrgh
Having a theory that relies on the multiverse is like making love to a beautiful married woman.
You wouldn’t be doing it if you thought there was a chance you were going to be found out.
The correct phrase in relation to this is “jumped the shark.” See http://www.youtube.com/watch?v=MpraJYnbVtE
Well, although it’s not April 1, has anyone considered the possibility that this is just a joke?
The problem with the joke hypothesis is that these authors have written quite a few papers already that aren’t that different. If it’s all a joke, they’re keeping it up for a remarkably long time.
@ Christian Takacs: Well if time ends there’s nothing stoping the person 3 billion years in the future from just walking over and telling you right now, right?
Quite honestly, I do object to long term “testible” predictions being made about anything 3.3 billion years into the future with any level of statistical accuracy worth mentioning. Please consider the very brief period of time humans have been observing the cosmos, and how many times over just the last 100 years the understanding of how this cosmos functions has changed.
To make such long term predictions with so little understanding of how gravity, light, space, and time function in this universe(not to mention as yet uknown variables) on such a large scale is pure hubris.
I believe that any math is still bound by logic; If you can’t observe and measure something, you certainly can not make accurate models or predictions .
The difficulty with many successful so called theoretical physicists today is their complete lack of an education in philosophy and the humanities. They were selected mainly on their ability to mimic the computational skills of their teachers, with no regard for real scientific enquiry, mind boggling egos and a complacency that would astound any previous generation of human being.
How many of them have ever worked in a real laboratory? Or in industry? Who would have thought it was possible to wander so far from genuine scientific prediction. And when the public talk of massive funding cuts, these same peole cry foul for the future of their societies. Good riddance to it.
The weird thing is that many “multiverse” papers like this don’t involve any computational skills beyond high-school level math. They consist of pages and pages of natural language text with a few trivial equations interspersed. The problem isn’t that the authors are showing off computational technique without addressing philosophical issues. They’re often addressing nothing but philosophical issues, but doing so in a naive and pre-scientific style.
5B years will not be enough time for black holes to evaporate from Hawking radiation, but this story claims that the radiation has already been observed in the lab!
Do you also object to predictions that our Sun will explode into a red giant in a similar time-frame? It is possible for a testable theory to make long-term predictions, so it is not enough to just argue that people will not live that long.
“All men are mortal” is tested by looking at deaths today and not finding any reason that those alive today are somehow exempt from the same processes. Likewise, the theory and observation is that “stars like our sun do whatever” and so the prediction is that our sun will also do whatever.
I am unable to agree with this statement. Can anyone who agrees with this statement please explain why it might be true?
Can I join the fun and make up a new theory of physics, too? My theory is that quantum mechanics and general relativity are irreconcilably incompatible. This will only become physically relevant at around the time when the first black hole evaporates, around 1060 years from now, at which point the universe will end.
Isn’t this a much better theory than the one we’re discussing? Eternal inflation is still a speculative idea, while both quantum mechanics and general relativity are extremely well tested experimentally.
But I guess a paper simply pointing out that there are serious problems with the probability measure for eternal inflation wouldn’t have gotten anywhere near as much press as this paper did.
I think you’re missing the innovative aspect of this paper (and its similar cousins). In old-style rationalist argumentation, when you show that your hypotheses lead to absurd conclusions, you have to abandon one of more of your hypotheses. The innovation here is to turn a bug into a feature: when your hypotheses lead to an absurd conclusion, you write a paper (and possibly issue a press release) proclaiming that you have made a surprising and counter-intuitive discovery.
As long as you’re careful to avoid dealing with anything experimentalists can check, and stick as much as possible to vague words, avoiding precision and equations, there’s no danger of running into too-obvious-to-ignore logical contradiction.
10^60. Phew. Thank goodness. You had me worried there for a minute.
Well, you know it is more fun to work in a system with inconsistent axioms, since then you can prove ANYTHING 🙂
Pingback: Not Physics « Calcutta Chronicles
I just want to remind to the people who mock Bousso (not the paper but the actual person) here and elsewhere that he is the author (together with Polchinski) of the celebrated paper “Quantization of Four-form Fluxes and Dynamical Neutralization of the Cosmological Constant”.
This paper contains the only explanation the human race has so far for the observed value of the CC.
In any case I think a little respect is in order here.
If that paper hadn’t been written by Bousso and Polchinski, I doubt it would be considered much of an explanation.
Of course, as an example of how one “explains” observations of fundamental significance in cosmology and astrophysics using M-theory, I suppose it’s fairly typical.
(As for “the only explanation”, that is also open to dispute, but never mind that.)
About it being “testable”: ignore the long time period and think about if the prediction was there was a 50% change time will end tomorrow. Assume that’s correct. Then half of the time, time keeps on going and we gain no additional information. The other half of the time, time ends — and we no longer exist to gain confirmation the theory was true!
I’m not mocking Bousso, just the paper…
One common reaction at the time to the Bousso-Polchinski (and later, KKLT and Susskind) string theory anthropic landscape explanation of the CC, was that if you took that seriously, you were giving up on conventional scientific research and this would unleash all sorts of nonsense. There’s now significant evidence for how that turned out, including this latest paper.
I didn’t know that they accepted science fiction at the ArXiv – this idea could be turned into a great Star Trek episode…which is about the only (potentially) good thing about it.
> The difficulty with many successful so called theoretical physicists today is their complete lack of an education in philosophy and the humanities. They were selected mainly on their ability to mimic the computational skills of their teachers, with no regard for real scientific enquiry, mind boggling egos and a complacency that would astound any previous generation of human being..
Hmmm…. tried reading any philosophy recently? Just last week, a philosopher on the Web was trying to explain to me how philosophers have proved time travel to be impossible. (No, I’m not shilling for time travel – I just question whether the philosophers have proved that it is impossible).
Kea also wrote:
> And when the public talk of massive funding cuts, these same peole cry foul for the future of their societies. Good riddance to it.
I think there may be a testable prediction there: when the public really finds out about all this, their reactions will be…?
Dave Miller in Sacramento
Pingback: עדכונים: מתמטיים | ניימן 3.0
Dave, it is agreed that no academic discipline is immune to the disease of the times, but that is no excuse to ignore the discipline entirely!
Peter sorry for the spam but have a look at Andy Strominger’s talk on string theory (at Harvard)
see the discussion around 50-55 mt. its a string theory report card
Interesting looking talk, it’s at
If I can find time soon to watch it, may write something about it here.
And this in from The Economist’s blog, commenting on Hawking’s scientific credentials:
One can’t fault Hawking for this bizarre assessmnent of “what is expected,” but it is something of an index of how confused the public is.
Kea wrote to me:
>Dave, it is agreed that no academic discipline is immune to the disease of the times, but that is no excuse to ignore the discipline entirely!
My reply to you disappeared, so I suppose Peter does not want a lengthy discussion on this.
So, I’ll simply say that philosophy has been doing for two millennia what the string theorists have been doing for just a few decades.
And, yes, that indeed is a good reason for ignoring their discipline entirely.
ah yes, i have been told in school that physics has two dangerous limits: zoology and philosophy. seems these guys hit one of the limits pretty hard.
“Hmmm…. tried reading any philosophy recently? Just last week, a philosopher on the Web was trying to explain to me how philosophers have proved time travel to be impossible.”
Yes, one philosopher is representative of the entire discipline. How very logical of you to also accept his argument at face value, and then to make further inferential claims about an entire discipline. Especially when said discipline has multiple sub-areas, and sub-sub areas. The philosophy of time itself is split up into multiple conceptual areas , had you gone beyond your limited sample space of one “philosopher on the web” (who I’d wager good money wasn’t a philosopher at all, though it’s up to you to provide proof, please link it if this was an open web conversation …) you might have encountered the different and varied argumentative positions for philosophers of time (many of whom disagree on whether time travel can or cannot exist, and some — mostly of the empirical bent — deny whether it is meaningful to talk of completely).
Sounds like you need to read some more philosophy, namely informal and formal logic.
(1) And before someone goes “LOLPHILOSOPHERSCAN’TAGREE,” let it be known that philosophy is primarily useful insofar as it keeps a philosophical inventory of arguments that are known to be useful, or useless (primarily the latter, especially for metaphysical concepts) and groups them together by category. In the case of philosophy of time, and its sub-areas, there are numerous conceptual schemes and arguments for-and-against different positions. So to say “philosophers as a whole reject time travel” shows an ignorant account of philosophy in general.