Why Stick to GR? A Closer Look at the Theory's Successes and Predictions

[greypilgrim]

Hi,

GR explains the perihelion precession of Mercury, the deflection of light and the gravitational redshift pretty accurately. However, predictions for larger scales seem to be quite wrong, and physicists introduce stuff like the cosmological constant, dark matter and energy.

Why are they still so sure that GR is correct and hence dark matter and energy must exist? Why is this more likely than GR being wrong and having to be replaced by a theory that doesn't need such inelegant corrections like dark matter and energy?

I'm a bit surprised that GR is celebrated as being incredibly successful while its predictions are just so wrong on cosmological scales so far.

 

[Dale]

Do you have a scientific source which describes any of these specific wrong predictions on cosmological scales that you are referring to?

 

[elusiveshame]

I'm no expert on GR, but wasn't the cosmological constant included in it from the start?

 

[Orodruin]

Why are they still so sure that GR is correct and hence dark matter and energy must exist? Why is this more likely than GR being wrong and having to be replaced by a theory that doesn't need such inelegant corrections like dark matter and energy?

There are other theories around that attempt at solving the same problems with modified gravity such as Modified Newtonian Dynamics (MoND). These theories tend to have some problems in relating to observations of colliding galaxies and galaxy clusters.

To get things straight, dark matter and dark energy are not corrections to GR. They are simply other forms of fluids which are allowed in GR, but behave differently from regular matter (well, dark matter behaves as matter as far as GR is concerned, it does not behave as matter when it comes to being luminous).

I'm no expert on GR, but wasn't the cosmological constant included in it from the start?

GR has always allowed for a cosmological constant. It makes no a priori prediction of the value of such a constant.

I'm a bit surprised that GR is celebrated as being incredibly successful while its predictions are just so wrong on cosmological scales so far.

The predictions you are talking about are based on your inherent assumption that normal matter is all there is. It is not an intrinsic prediction of GR, which is perfectly compatible with the large distance observations with a component of dark energy and an additional dark matter component.

The fact is that GR has made such precise predictions that any theory you can come up with is going to have to reproduce these predictions to even be considered.

 

[elusiveshame]

GR has always allowed for a cosmological constant. It makes no a priori prediction of the value of such a constant.

Thank you for the clarification :)

 

[greypilgrim]

Do you have a scientific source which describes any of these specific wrong predictions on cosmological scales that you are referring to?

I mean all the observations that we ascribe to the existence of dark matter and energy, such as the shape of galaxies. It seems pretty far-fetched to me to postulate that the theory is correct, but we just haven't accounted for 95% of the energy in the universe which just happens to be invisible

Also makes me wonder about the falsifiability of GR. For any experiment disagreeing with standard GR we could just introduce invisible pink matter and tune its properties until GR + pink matter agree with the observations.

 

[Orodruin]

For any experiment disagreeing with standard GR we could just introduce invisible pink matter and tune its properties until GR + pink matter agree with the observations.

No, this is not true. There are some very specific statements within GR on how strange the components of the Universe may behave, encoded in the GR equivalent of energy and momentum conservation ##\nabla_\mu T^{\mu\nu} = 0##.

Regarding the falsifiability, GR has already passed a large number of tests where it could have been falsified, such as gravitational lensing, gravitational time dilation, frame dragging, etc. Within GR, there is no way of tweaking those results to fit observations. If you find it unappealing, nobody will care unless you can figure out a model which predicts things better (or at least as well but with fewer assumptions) - this is the scientific method at work. Many people find quantum mechanics unappealing as well, but why should the way the Universe behaves be appealing (which is anyway a subjective concept)? Richard Dawkins gave an interesting talk on how we should not be surprised by finding strange things when looking at scales with which we are not intuitively familiar http://www.ted.com/talks/richard_dawkins_on_our_queer_universe

 

[Dale]

I mean all the observations that we ascribe to the existence of dark matter and energy, such as the shape of galaxies. It seems pretty far-fetched to me to postulate that the theory is correct, but we just haven't accounted for 95% of the energy in the universe which just happens to be invisible

Why? What fraction do you think we have accounted for, and why that fraction?

The problem is that "farfetchedness" isn't a scientific criterion, and scientific theories are judged on scientific criteria. There are essentially two scientific criteria. The first is consistency with observation, and the second is simplicity (few free parameters).

Many alternative theories have been proposed and others are in development. However, so far all alternative theories have either agreed with GR or disagreed with experiment. Of those that agree with GR, all of them contain additional free parameters. So by scientific criteria there simply isn't a better alternative. Yet.

 

[wabbit]

Another viewpoint, debatable but perhaps better grounded, is to ask "why is the cosmological constant so small ? It would be far more natural for it to be many orders of magnitude larger". If it were though, dark energy would account for 99.99... % of the energy of the universe instead of 70 %. We would probably not be here to discuss it, but I think this illustrates that "far-fetched" can be depend on how you look at it.

Also, GR predicts that this percentage generically increases as the universe ages and ultimately approaches 100 %, so its current value is in a way just a clock telling us our universe is middle-aged : )

 

[Nugatory]

The short answer to the question posed in the thread title is "Because it's worked better than anything else so far and no one has a better idea". The longer answer...

Why are they still so sure that GR is correct and hence dark matter and energy must exist? Why is this more likely than GR being wrong and having to be replaced by a theory that doesn't need such inelegant corrections like dark matter and energy?

In cosmology we apply the Einstein field equations to a hypothesized distribution of matter and energy, so a bad prediction could indicate that the field equations don't apply to cosmological conditions or that our hypothesized distribution of matter and energy isn't right. The field equations have never let us down when we give them known-good inputs and our confidence in the hypothesized distribution of matter and energy is rather low, so it is natural to interpret the failed predictions as tests of the hypothesis about the distribution of matter and energy instead of the applicability of GR.

You are right that this is an assumption so can be challenged (although to be taken seriously the challenge has to include a better idea or an unambiguous falsification based on high-confidence experimental conditions and observations). We may end up discovering some other theory that reduces to GR at less than cosmological scales (Orodruin has mentioned some candidates elsewhere in this thread); we already expect to find corrections to GR at the opposite end of the spectrum under conditions of very high energy and very small size as in the the singularity at the center of a black hole. However, even if this happens it would be a mistake to conclude that GR is "wrong"; http://chem.tufts.edu/AnswersInScience/RelativityofWrong.htm, that's not how successful theories are replaced.

I mean all the observations that we ascribe to the existence of dark matter and energy, such as the shape of galaxies. It seems pretty far-fetched to me to postulate that the theory is correct, but we just haven't accounted for 95% of the energy in the universe which just happens to be invisible.

It's not that far-fetched if you consider that essentially all of the astronomical phenomena we study today were invisible to the visible-light astronomers of a few generations back. The people who are considering the hypothesis that with current technology 95% of the energy in the universe can be detected only through its gravitational effects have a pretty good understanding of the limitations of that technology.

Also makes me wonder about the falsifiability of GR. For any experiment disagreeing with standard GR we could just introduce invisible pink matter and tune its properties until GR + pink matter agree with the observations.

Except that we don't have unlimited freedom to tune its properties. We have to choose properties such that when we plug this hypothetical pink matter into the field equations, what comes out matches observation - and the inability to do that would be a successful falsification.

 

[pervect]

As others have mentioned, there are alternate peer-reviewed published theories of gravity other than GR which attempt to do away with dark matter. The only one I'm at all familiar with (and I'm not terribly familiar with it even so) is MOND. WIki has an article on it, http://en.wikipedia.org/w/index.php?title=Modified_Newtonian_dynamics&oldid=665554848 that seems sensible. According to the Wiki article while Mond and it's newer relativistic variants have had some success in explaining the things it was created to explain, it has not had any success in explaining other things. This is usually regarded as a bad sign, a good theory is one that not only explains what it was created to explain, but either predicts new phenomenon that are confirmed by testing or explains other previously unexplained results that it was not created to explain. Explaining some previously unexplained results at the expense of not-explaining other known results is not a promising start to a theory.

GR is still regarded as the "gold standard" of gravity, for good reasons. It's made numerous predictions of phenomenon that were not expected or even measured at the time of its creation, starting with the original light bending experiments, continuing on with many other solar system tests (such as the Shapiro delay, the Cassini probe being the most accurate test of this kind), gravitational time dilation (easily measurable and so important that we need to compensate our atomic clocks for height in order to create a precision coordinate time), and most recently gravity probe B's detection of frame dragging. It's also explains the previously known but previously unexplained precession of Mercury's orbit. GR almost predicted the Hubble expansion, but did not, because the original theory included a cosmological constant (the same one you objected to). The cosmological constant was done away with for a time, but wound up coming back as more precise cosmological data was taken. GR also has had a great deal of success in explaining other cosmological observations of the anisotropy of the Cosmic Microwave Background radiation - look up CMB and the WMAP (Wilkinson Microwae Anisotropy probe). My understanding (from wiki) is that MOND-type theories have difficulties explaining these cosmological results.

The current consensus view is that dark matter is probably real, but everyone wants more data and a better explanation of exactly what it is. While the area is one of active research, dark matter is on a much more solid experimental basis than "invisible pink elephants". Additionally, to the best of my knowledge, the candidate theories that attempt to eliminate dark matter would not only requiring falsifying General Relativity, but also require changes to Newtonian gravity as well.

 

[russ_watters]

To get things straight, dark matter and dark energy are not corrections to GR. They are simply other forms of fluids which are allowed in GR, but behave differently from regular matter (well, dark matter behaves as matter as far as GR is concerned, it does not behave as matter when it comes to being luminous).

GR has always allowed for a cosmological constant. It makes no a priori prediction of the value of such a constant.

The predictions you are talking about are based on your inherent assumption that normal matter is all there is. It is not an intrinsic prediction of GR, which is perfectly compatible with the large distance observations with a component of dark energy and an additional dark matter component.

The fact is that GR has made such precise predictions that any theory you can come up with is going to have to reproduce these predictions to even be considered.

I'm not sure how we'll that was absorbed, so let me reiterate in a somewhat different way:

GR sets up a lot of the rules that govern how the universe operates, but in order to actually describe it, you need input data and proportionality constants, which are tough to gather since the systems we are trying to examine are billions of light years away.

The objection is like saying GR and Newton''s theories fail because they can't accurately predict the mass and behavior of the Pluto system. No, they surely can, we just lack the input data required to refine the model because we've never been there (an issue to be resolved shortly).

So you are confusing the theory (which as far as we know, works) with the model of the system (which requires additional data to complete).

 

[Orodruin]

Let me also draw a parallel to the discovery of Neptune (http://en.wikipedia.org/wiki/Discovery_of_Neptune). Newton's law of gravity was well established but did not accurately predict the orbit of Uranus. The discrepancy could be explained by the "dark" matter of that time, a planet which was too dim to be seen by current technology and so Neptune's existence was predicted. Eventually, this "dark" matter was found using better instrumentation. Newton's theory of gravity was not in serious peril as confidence in it was high after several accurate predictions. The discovery of Neptune was consequently added to the pile of successful predictions of Newtonian gravity. I would say it is not too much of a surprise if this history repeats itself with our dark matter playing the role of Neptune.

 

[Dutchr]

I would be greatly surprised if any theory we currently have matches nature exactly. I am probably sure that most the theories we currently have will be replaced with something a little bit better in the future. However, we still use Newton's theories even till this day. When a theory is elevated as far as Newton's theory of gravity or Einstein's GR science never gets it "wrong" as those theories are well tested. Each new theory gets more refined and explains more than its predecessor. Nevertheless, any scientist should realize that no theory can ever be elevated to absolute fact.

We have to use the best, simplest, and most accurate theories as that's all we have.

Dark matter may be real or it may be a false prediction. Let the experiments decide.