Is Gravity Weakening as the Universe Expands?

[Naty1]

In 1939 the famous British mathematician and physicst P. A M. Dirac developed a theory in which gravity is slowly weakening as the universe expands and its density thins.

Anybody know if there is any validity to this and whether any new theoretical work has been done?

The source references a 1976 article in Scientific American "Is the gravitational constant constant". After Fritz Zwicky's proposal in the early 1930's nobody did much on dark matter until Vera Rubin's work around 1970, so maybe that was part of the article...

The source goes on to imply that things might change ever so slightly, like plate tectonics,large bodies in the solar system might expand (slightly) and our own sun might have been slightly hotter and denser in the past...contributing to the previous widespread tropical atmosphere..

RELATIVITY SIMPLY EXPLAINED, Martine Gardner, 1997

 

[twofish-quant]

Anybody know if there is any validity to this and whether any new theoretical work has been done?

It's an observational question, and the observations say that G isn't changing much...

http://www.rssd.esa.int/SA-general/Projects/GAIA_files/LATEX2HTML/node143.html

The source goes on to imply that things might change ever so slightly, like plate tectonics,large bodies in the solar system might expand (slightly) and our own sun might have been slightly hotter and denser in the past...contributing to the previous widespread tropical atmosphere..

It's one of those ideas that wasn't totally nutty given what we knew in 1930, but we know more now.

 

[Naty1]

By source I meant the author who was writing in 1997...assuming the book was updated from two earlier editions...I AM really asking a theoretical question...we can't even measure gravitational waves so far, so a change as tiny as is implied by the author is likely even further beyond our observational capabilities...

 

[bapowell]

By source I meant the author who was writing in 1997...assuming the book was updated from two earlier editions...I AM really asking a theoretical question...we can't even measure gravitational waves so far, so a change as tiny as is implied by the author is likely even further beyond our observational capabilities...

What does measuring variations in G over time have to do with detecting gravitational waves?? The CMB can be used to constrain the value of G at recombination relative to its value today. See http://arxiv.org/abs/0905.1808" : these guys find an allowed variation at 1-sigma of about 10%.

 

[twofish-quant]

I don't think that book was updated. Looking at google, the book was originally published in 1962, and in 1962, you might be able to get away with arguing that gravity is weakening, but there is just no way could argue that in 1997.

 

[twofish-quant]

we can't even measure gravitational waves so far

Actually, we can measure gravitational waves just fine. It's just that it appears that there is nothing producing gravitational waves that are strong enough to hit detection limits.

so a change as tiny as is implied by the author is likely even further beyond our observational capabilities...

Nope. Changes in G have nothing to do with gravitational waves, and even small changes in G would have huge, huge impact.

 

[Wallace]

According to GR, you can't simply make G = G(t) and plug and play your existing G=constant equations. To do so would violate Lorentz invariance, which lies at the heart of the concept of a 4D spacetime (essentially says that time is just another dimension, even if is behaves a little differently to the spatial dimensions).

The paper mentioned in post #4 has come under a fair amount of criticism because they ignore this fact. Essentially they just replace G with G(t) in the usual cosmo analysis codes and published the results, unfortunately that's not physically valid. This imposes a physical law which would only hold in one co-ordinate system, which is not permitted under GR (or essentially any other post-Newtonian theory of Gravity).

What you need to do (and what has been attempted) is much trickier, and that is to add extra fields to GR that look like they make a G -> G(t) change in the homogeneous case, however when you do perturbation theory you find that G effectively ends up varying in both time and space, but you none the less have a consistent set of perturbation equations that you could use in your cosmo analysis codes. What that work did was use the G constant solutions and then add a G(t), rather than re-deriving the perturbation equations consistent with a new physical model. The differences between the approaches are enough to matter for this kind of analysis.

I'd also argue that we have effectively detected gravitational waves in measurements of binary pulsar spin down rates, but that is admittedly not a 'direct' detection. There are also hints in CMB polarisation measurements that point to primordial gravitational radiation, although that is a much less convincing observation that the Pulsars at this point.

 

[bapowell]

There are also hints in CMB polarisation measurements that point to primordial gravitational radiation, although that is a much less convincing observation that the Pulsars at this point.

Really?? Are you referring to the B-mode? I wasn't aware that anything had been seen to date...

 

[Wallace]

Yeah I'm referring to B-modes in the CMB. My understanding is that there are some hints pointing to those in what has been seen, but they are really just hints, not anything solid yet. I'm not 100% sure on that though, so I could be confusing some projections of future measurements with current data!

 

[Chalnoth]

Yeah I'm referring to B-modes in the CMB. My understanding is that there are some hints pointing to those in what has been seen, but they are really just hints, not anything solid yet. I'm not 100% sure on that though, so I could be confusing some projections of future measurements with current data!

Yeah, there has been as yet no direct detection of B-mode polarization in the CMB. Unfortunately, the level of B-mode polarization depends upon the nature of inflation, so we don't know how difficult it will be to detect.

 

[Wallace]

Hmm, I thought that there were definitely B-modes seen, but gravitational lensing of the CMB also induces B-modes, and the extraction of the 'de-lensed' CMB (by referring to galaxy redshift surveys to model the lensing) showed residual B-modes that could be primordial or could just be noise from the extraction, but the whole process is not yet good enough (due to limitations in the data) to know much for sure about the primordial part yet with any confidence.

I'm just a theory guy though, so I'm not sure exactly the current state of the data, that's just the impression I have from talks I probably slept most of the way through...

 

[Chalnoth]

Hmm, I thought that there were definitely B-modes seen, but gravitational lensing of the CMB also induces B-modes, and the extraction of the 'de-lensed' CMB (by referring to galaxy redshift surveys to model the lensing) showed residual B-modes that could be primordial or could just be noise from the extraction, but the whole process is not yet good enough (due to limitations in the data) to know much for sure about the primordial part yet with any confidence.

I'm just a theory guy though, so I'm not sure exactly the current state of the data, that's just the impression I have from talks I probably slept most of the way through...

Nope, according to the guy who I share an office with, who is directly involved in one of the experiments intending to detect B-modes states that the current best upper limits on them (even from lensing) are from BICEP, with QUAD coming in a close second. No actual detection so far, just upper limits.

 

[Chronos]

The consensus is Hulse nailed the case for gravitational waves in binary neutron star studies.

 

[Naty1]

Chanoth:

Yeah, there has been as yet no direct detection of B-mode polarization in the CMB

That was the consensus on physics forums in another thread within the last six mos or so.

Wallace

According to GR, you can't simply make G = G(t) and plug and play your existing G=constant equations. To do so would violate Lorentz invariance

Interesting point!...had never thought about that...yet I don't think G was constant early on in the universe assuming a bang type start...is that correct and we think it has been stable since?...

 

[Wallace]

As far as GR goes, G is a constant at all times (and places!). There is nothing about current early universe understanding that involves a changing G. That being said, there is obviously a lot we don't understand about the early Universe, and we don't exactly know how far back GR is valid. However, whenever GR is a valid theory, you can't simply make G -> G(t).

 

[yogi]

Most efforts to determine variable G involve long term radar ranging experiments that show the orbits do not vary much during the sampling period - but these conclusions do not measure G separately, but rather some MG product. If inertial mass of a local moon depends upon all other cosmic mass within the Hubble sphere a la Mach, then any conclusions about the temporal invariance of G would require an experiment that measures G alone exclusive of M. I am not aware of any experiments that can single out G (anyone know of any, I would like a reference if anyone has one). If the cosmic energy is not conserved, or if it is created during expansion, then if Mach's principle applies - the measurement of MG invariance guarantees the variance of G.

 

[Chalnoth]

Most efforts to determine variable G involve long term radar ranging experiments that show the orbits do not vary much during the sampling period - but these conclusions do not measure G separately, but rather some MG product. If inertial mass of a local moon depends upon all other cosmic mass within the Hubble sphere a la Mach, then any conclusions about the temporal invariance of G would require an experiment that measures G alone exclusive of M. I am not aware of any experiments that can single out G (anyone know of any, I would like a reference if anyone has one). If the cosmic energy is not conserved, or if it is created during expansion, then if Mach's principle applies - the measurement of MG invariance guarantees the variance of G.

Because G has units, it's not possible to measure it independently.

 

[yogi]

G has units of volumetric acceleration per unit mass - if we know the acceleration of the observable universe in anyone direction (e.g., if it is c^2/R) then do we not have an independent measure of G?

 

[Chalnoth]

G has units of volumetric acceleration per unit mass - if we know the acceleration of the observable universe in anyone direction (e.g., if it is c^2/R) then do we not have an independent measure of G?

You can always change such an acceleration by changing your coordinates.

 

[yogi]

Every non-accelerating observer will measure the same value of cosmological acceleration at the Hubble radius R

 

[Chalnoth]

Every non-accelerating observer will measure the same value of cosmological acceleration at the Hubble radius R

Except that doesn't merely depend upon G, but also the energy density of the various components of our universe.