Evidence of the Universe from Before the Big Bang? (Synopsis)

Despite its name, the big bang theory is not really a theory of a bang at all. It is really only a theory of the aftermath of a bang.” -Alan Guth

So, you've heard the big news by now, no doubt? Primordial B-modes have been detected in the polarization of the cosmic microwave background!

Images credit: Seljak & Zaldarriaga (L), Wayne Hu (R), via http://cosmology.berkeley.edu/~yuki/CMBpol/CMBpol.htm. Images credit: Seljak & Zaldarriaga (L), Wayne Hu (R), via http://cosmology.berkeley.edu/~yuki/CMBpol/CMBpol.htm.

But how robust is this? Will it hold up under scrutiny? And is it statistically significant enough to call a "discovery" just yet?

And finally, whether it does or it doesn't, what does it tell us about the origin of our Universe?

Go get all these question answered (and more) at the new Starts With A Bang on Medium!

More like this

Why is this being reported at the 5-sigma level if, as you state, it isn't?

I was wondering the same thing. For example, from their FAQ:
"We have detected B-mode polarization at precisely the angular scales where the inflationary signal is expected to peak with very high significance (> 5 sigma). We have extensively studied possible contamination from instrumental effects and feel confident we can limit them to much smaller than the observed signal. Inflationary gravitational waves appear to be by far the most likely explanation for the signal we see. "

By Adam Kamp (not verified) on 17 Mar 2014 #permalink

I'm also curious about the idea that inflation occurred before the big bang. As I understand it, Big Bang theory incorporates inflation models into it to explain away initial problems the original theory could not account for.

From the article you've referenced (p. 21):

"This calculation results in rejection of the hypothesis that there is no lensing B-modes with 4.7[sigma] confidence for a normal distribution."

What am I missing here?


Please, provide more details about this whole issue, and in layman's terms, if you will:) All this talk about small-scale and large-scale inflation, about the r parameter etc. gives me no insight, only big headache, as there was (I think) no talk about different types of inflation here up until now.

What each type of inflation entails? What does it tell us about our past? Does it tell us anything about the present (in terms of us living in a multiverse?)? Does it tell us anything about the future? And finally, when do you expect the data to be either confirmed (5-sigma significance)?


The lensing signal is not expected to peak at large angular scales (low L) modes like the gravitational wave signature is expected to.

This is a slight misinterpretation of the data. While lensing is the #1 foreground for primordial gravitational wave B-modes, it's expected contribution in the BICEP(2) angular scales/range is very low.

BICEP only detects the lensing signal at 2.7sig because their range of angular scales does not go to small enough resolution to resolve the peak of the lensing signal.

Conversely, you can see that the lensing signal theoretical curve is VERY low at the peak of the graviational wave BB curve. You would have to significantly modify GR to allow lensing to explain this signal.

Look here in this plot: https://d262ilb51hltx0.cloudfront.net/max/800/1*p9MwQDAKvehtxTJSFyZHJQ…

If you covered up the dashed line, the data points, especially at low L, are WAY MORE than just 3 sigma away from the solid red line. The low L modes are the peak of the GW signal, and it fits nicely with that dash curve, but isn't even touched by the solid curve.

As comment #3 alluded to, I too am curious about the chronology of the Big Bang. In this blog post, as well as others, you refer to the Inflationary Epoch as the era prior to the Big Bang. Other sources describe it the other way around. Even you Blog title, 'Starts With a Bang,' implies that the Big Bang occured before Inflation. Any clarification on this point would be appreciated.

Whoa! I didn't like this:

"In these regions — regions where inflation ends — we get a Universe, and one that’s much bigger than just the part observable to us. That’s the idea of a Multiverse, and why we think we almost certainly live in one".

Let's keep a sense of proportion here. What's been detected is some B-mode polarization of the CMB. Associated with this, are some big claims, and some media hype. If it all starts turning into damn statistics and 5σ "evidence" for the multiverse, cosmology will be lost as a serious science.

By John Duffield (not verified) on 17 Mar 2014 #permalink

@ 9

James, the idea that we live in a multiverse is not related to this discovery (in that it's not validated by it or not). Search this site for other articles about it.


Sorry, I meant 'John'. Not exacly sure why I type 'James'. No offence meant.


I'm not sure about when you say:
"In fact, it’s these special deformations of spacetime, on specific scales, that will stretch the wavelengths of light in a very particular fashion as they travel from the Big Bang to our eye."
As I understand it, gravitational waves cause temperature fluctuations in the early (pre-recombination) universe by applying red- and blueshifts to photons in the plasma. Those temperature fluctuations then lead to the conditions needed to generate polarisation in those photons, and so the polarisation signal is imprinted in the run up to recombination, and locked in then.
After that you get the lensing foreground signals applied of course, but the primordial signature is in the CMB at the time it starts its journey to us - that isn't an effect as they travel from the surface of last scattering to our eye.

Over at The Guardian (U.K.) newspaper's website, a report
( http://www.theguardian.com/science/2014/mar/18/stephen-hawking-gravitat… ) says,

"Stephen Hawking has claimed victory in a bet with a fellow scientist over the discovery of primordial gravitational waves, ripples in the structure of space-time from the birth of the universe.

The Cambridge cosmologist bet Neil Turok, director of the Perimeter Institute in Canada, that gravitational waves from the first fleeting moments after the big bang would be detected.

"Speaking on BBC Radio 4's Today programme, Hawking said the discovery of gravitational waves, announced on Monday by researchers at the Harvard-Smithsonian Centre for Astrophysics, disproves Turok's theory that the universe cycles endlessly from one big bang to another.

"If confirmed by other groups, the discovery would count as the strongest evidence yet for cosmic inflation, a theory which says that the universe went through a period of extremely rapid expansion soon after the big bang. The theory explains why the universe looks almost the same in every direction.

"It is another confirmation of inflation," Hawking told the Today programme. "It also means I win a bet with Neil Turok, director of the Perimeter Institute in Canada, for cyclic universe theory predicts no gravitational waves from the early universe."

Can someone offer or point to a layman's explanation for why that is supposed so? i.e Why are primordial gravitational waves and a cyclical big-bang theory necessarily mutually exclusive?

By proximity1 (not verified) on 18 Mar 2014 #permalink

According to the actual BICEP2 paper, http://arxiv.org/pdf/1403.3985v1.pdf, the 2.7 sigma figure refers to "de-
tection of lensing in the BICEP2 BB autospectrum." This is not the important result of the paper.

BICEP2 was specifically not looking for evidence of gravitational lensing. They were looking for evidence of primordial gravitational waves leaving a specific imprint on the CMB polarization. Thus, the relevant and important result in the paper is summed up by the following sentence:

"This excess represents a 5.2 sigma excursion from the base lensed-LambdaCDM model."

That means there is a large signal above and beyond what is expected due to lensing, which represents rather strong evidence (5 sigma evidence, which in the particle physics community is enough to lable something a "detection") that something else interesting is going on. The most likely explanation is tensor perturbations from gravitational waves created at around the time of the big bang itself.

Sorry, there was a typo in my last post. It should say label, not lable.

I think Kyle at #7 is correct regarding the interpretation of the lensing detection by BICEP2. Its angular resolution is of order 15 arcminutes, which isn't that great for probing the lensing B-modes, so the 2.7-sigma detection is just of the lensing stuff itself by BICEP2. I think it's an extension of the combined Polarbear/BICEP2 measurements of the lensing foreground to ell = 40 that sets the lower range of the reported r = 0.20 +0.07/-0.05. That means it's a 4-sigma detection of a non-zero tensor-to-scalar ratio.

By Tom Renbarger (not verified) on 18 Mar 2014 #permalink

Also, in the interests of full disclosure, I should have mentioned that I was a postdoc in Brian Keating’s lab from 2004-11, and he's one of the co-I's for BICEP2 (and co-PI of Polarbear), but I’ve been out of the observational cosmology field for three years now.

By Tom Renbarger (not verified) on 18 Mar 2014 #permalink

"Phenomena" is the plural of "phenomenon"!#! One phenomenon, , two or more phenomena. Thanks for the opportunity to vent.

By Dan MacDuff (not verified) on 18 Mar 2014 #permalink

I have been concerned for some time about the so-called 5 sigma concept that is used to validate signal detection. To say that a signal has only one chance in 10 million (or 10 thousand, or whatever level of certainty is the "in" thing), of being false requires a knowledge of the underlying noise distribution to a degree which is very difficult to attain. I spent many years working to detect elusive signals in difficult noise backgrounds and I know from experience that nature rarely (never in my experience) produces Gaussian background distributions in the "tails" which dominate the detection problem. It is usually true that events which are 10E-7 events in a Gaussian world are actually only 10E-5 events in the real world (or worse). Do you know whether the background noise statistics for this paper are assumed to be Gaussian, or have they accumulated enough noise data to actually characterize the noise environment to a level sufficient to support their assertions (it takes a LOT of noise data, btw).

Noted Dan. I'm afraid whenever I hear 5σ nowadays I feel sceptical already. There's an awful lot of assumption with all this. What's actually been detected is some B-mode polarization of the CMB. The CMB dates from the epoch of recombination. It doesn't date from the inflationary epoch. There's just too much inference for my liking.

gaijin: no probs re the name. Be alert for the multiverse thing, because I tell you, that's where this is going.

By John Duffield (not verified) on 18 Mar 2014 #permalink

@ John

the fluctuations in CMB date from inflationary period

By Sinisa Lazarek (not verified) on 19 Mar 2014 #permalink

Echoing comments #3 and #8, I am too quite a bit confused by the notion that inflation did in fact happen "before the Big Bang" as Ethan seems to suggest.
All sources I can find state the contrary, and indeed I too thought the BB was meant as the starting event that gave rise to our universe, that is from time 0 onwards (or maybe, from 1 Planck time onwards).

Ethan, what exactly is the timeline we're discussing here? Wikipedia says about the inflationary epoch: "ending 10^–32 second after the Big Bang". I feel you're making some point I'm completely missing. Could you clarify that?

I am going to write a magnum opus on cosmic inflation explaining all of this. ALL of it!

Stay tuned. It takes a while to do it right.