Gravity and relativity.

March 28, 2006 2:47 AM

If General Relativity is accepted as true, why do Physicists talk about (and look for) Gravitons?

My confusion here is that the two approaches seem to be mutually exclusive - either gravity is a function of 'warped space-time', or it is a conventional force mediated by a particle. Can it be both? If so, isn't one explanation redundant?

If not, can I take it that researchers who talk about Gravitons, as in this bizzare press release, don't accept General Relativity?

A lot of questions here, I guess I'm really asking for a laymans account of how both approaches could be reconciled. Any good explanatory links as to the current thinking about gravity would also be appreciated.

posted by grahamwell to science & nature (14 comments total)

No theory is "true"; that's not how the system works.

And in the case of GR, it's actually known that it will have to be replaced, or so I understand it. (The magic words are "Bell's Inequality".)

The reason that physicists are looking for a quantum theory of gravitation is that they want all forces to be the same, with each mediated by a particle. As it currently stands, three forces are mediated by particles and the forth one is an aspect of the geometry of the universe, a distinctly inelegant result.

Of course, the universe didn't promise us it would be elegant, but physicists keep hoping it will end up that way. Fact is, I don't think there's really a lot more to it than that.
posted by Steven C. Den Beste at 3:00 AM on March 28, 2006

General Relativity is accepted as useful, not true.

Gravitons are a prediction from the Standard Model of particle physics, a discipline largely based on quantum mechanics. Quantum mechanics is also accepted as useful, not true.

In areas of inquiry where both GR and QM seem to be needed - such as working out just what is going on inside the event horizons of black holes, or what went on around the Big Bang - it is indeed often extremely difficult to decide which approach to use; as you note, the underlying assumptions of each seem to be incompatible.

But in the problem domains they were originally developed for - GR for questions about largescale behaviour, QM for questions about very small-scale behaviour - it's quite clear which set of models to use, and the results are in very good accordance with observational experience.

At present there are no respectable methods for reconciling the two approaches, so a layman's account is still some way off.
posted by flabdablet at 3:05 AM on March 28, 2006

Hmm. Fair enough, let me put it another way, would the discovery of a Graviton settle the matter and require a root and branch revision of General Relativity, specifically the supposed causal agent (space-time geometry) that is invoked to explain gravity?

And as a corollary, do enthusiasts for the relativistic explanation of gravity think that the particle hunters are wasting their time?
posted by grahamwell at 3:16 AM on March 28, 2006

I expect not; no more than the development of GR itself required a root-and-branch revision of Newtonian mechanics - which, though it contains embedded assumptions that are now known not to be universally justifiable, continues to work well enough to do stuff like design cars and plan space missions. Its assumptions are flawed, but there's a huge problem space where those flaws make no significant difference.

It pays to be very careful, when thinking about this stuff, not to mistake the map for the territory. You speak casually of space-time geometry being a "causal agent" for gravity; in fact, spacetime, mass, energy, gravity and the geometrical mathematics that connect them are all theoretical constructs invented to model certain observable phenomena. They all go hand in hand, and none of them can be said to cause any of the others.

Saying all this another way: though there are undoubtedly parts of the territory (reality) where our present maps (theoretical frameworks) aren't much use, those are parts that are hard to get to anyway - and once we do have maps for those, we may well end up using the old maps for the well-known parts anyway, just because they're well tested and known to work.

Extending the metaphor: you don't throw away your street directory just because you find out how to get around certain parts of the city by spelunking drains, and you don't measure your travel distances in millimetres in order to work out how long it will take to get to the office :)

I'm sure there are indeed GR enthusiasts who are confident that gravitons will not be found. I suspect, though, that most of them are not working physicists - rather, they will be lay people who treat science more like a religion than like the always-provisional search for useful models that it actually is.
posted by flabdablet at 4:23 AM on March 28, 2006

As the others say, theories in science are not accepted as 'true'. They earn their keep by making accurate predictions about the physical world.

GR and QM make predictions about the world at totally different scales. You can't use GR gravitation to predict what will happen to a bunch of subatomic particles and you can't use QM to predict the motion of a planet.

Worse, no-one has come up with a way of linking the two theories (the so-called 'unified theory' which is the Holy Grail of physics).

Gravitons offer a possible way of linking the two theories, since if particles could affect one another by exchanging gravitons in the same way that they exchange energy and information by exchanging photons, we could explain GR in terms of QM.

There is no contradiction even then, because GR does not explain *why* gravity is as it is, only *how* it is as it is. We learn from einstein that mass curves spacetime, but not how they do it, or indeed what spacetime is.
posted by unSane at 4:25 AM on March 28, 2006

No practicing physicist doubts that general relativity is a correct description of gravitational physics, provided that one considers only long distances and small spacetime curvatures.

Now, when one is given a "classical" (non-quantum) theory such as general relativity, there is a standard procedure for working out the "semiclassical" physics -- this is a sort of first approximation to the full quantum theory. Applying this procedure to general relativity, one arrives quickly at the prediction that there should be gravitons. One thinks of these gravitons as representing fluctuations of the space-time geometry -- the first-order way of visualizing them is just to think of a little wiggle in the fabric of the universe, which propagates through space and time. Most physicists today would say that this shouldn't be thought of as a revision of general relativity, but rather as a sort of minimal consequence of general relativity and quantum mechanics together.

Let me emphasize that the existence of gravitons is really a prediction of semiclassical physics, as reflected in the fact that I roughly described them above as fluctuations of the spacetime geometry, a description which makes reference to the classical general relativity picture. This semiclassical statement doesn't require the full strength of a theory of quantum gravity. That is why we can predict the existence of gravitons despite the fact that we notoriously don't have a complete theory of quantum gravity yet!

(Of course, until someone sees a graviton, it's always possible that there is something wrong with the standard reasoning.)
posted by em at 4:43 AM on March 28, 2006

as em says, it basically comes down to "hey, we have particles for everything else, so we need 'em for gravity".

that doesn't mean that GR is terribly wrong, or that quantum mechanics is terriby wrong, but that both are partial views of some more general underlying theory that unites the two.

it's like the blind men looking at the elephant (or whatever the story is; one feels the trunk and says it's a snake, etc.) - different parts of something not yet completely understood appear totally distinct. but when, hopefully, we have a better, more unified understanding of how the world works we'll probably end up with a theory (model) that at small scales agrees very nicely with models that talk about particles and waves (including gravitons) and which at large scales agrees very nicely with a model that talks about curvature of space and mass/energy.

think about how general relativity "replaced" newtonian dynamics. for most problems we still use newton's laws (and kelper's laws), because they're a good enough model for the case at hand.

other comments - bell's equality is more about "philosophical problems" with quantum mechanics that general relativity and there's currently an attempt to approach quantum mechanics as a kind of derivative of information theory which, if successful, might "explain" that.

and are they wasting their time? well, it's their time. it's perhaps more important to ask whether they are wasting money; that comes down to what the taxpayers want - in discussions i've had here americans seem quite happy to keep funding this kind of work. or perhaps you are asking - are they doing something stupid? and the answer is no, it makes sense within the context of physics reasearch.
posted by andrew cooke at 5:12 AM on March 28, 2006

so they're not mutually exclusive - you could have a model where matter/energy "warps" spacetime through a process that we model with particles. i have no idea how, but then i'm not expecting to win a nobel prize.
posted by andrew cooke at 5:15 AM on March 28, 2006

Also bear in mind that particles are abstractions. They're more your collection of numbers and rules than your tiny little billiard ball.

Physical theories are not for telling us what the world is like; they're for helping us predict how assorted bits of it will behave. At the kinds of extreme scales where Newtonian mechanics breaks down (i.e. starts making incorrect predictions), the world isn't like anything but itself, and our intuitions about it are misleading as often as not.

The tendency to see the word "particle" and picture a little billiard ball is what makes it hard for people to understand how light, for example, can be both a wave and a particle. The way to understand stuff like that is to remember that in the context of physical theory, a wave is just a mathematical model with a bunch of numbers attached to it; so is a particle.

Light itself is what it is. The quantum model we use to predict the behavior of light is neither a pure wave model nor a pure particle model, but a more complex model with both kinds of numbers attached.

Gravitons are that kind of thing too. So far, nobody has been able to point to any part of the real world and say with confidence "that's a graviton!" so it seems that gravitons are an abstraction in search of something to abstract; but there are many respectable physicists who fully expect those map entries to correspond usefully to a bit of territory any time soon.
posted by flabdablet at 5:45 AM on March 28, 2006

Scientific theories are not static ideas or finalized solutions. They are mathematical models describing certain known physical behaviors. Science is an iterative process and is constantly evolving based on new discoveries and changing perceptions of the world. As we make new discoveries with higher precision, scientists are forced to revisit old ideas and see if they remain true to the new conditions. Usually some minor tweaking is needed. Occasionally a total revamping needs to be done.
posted by JJ86 at 6:06 AM on March 28, 2006

Flabdabet and em have covered this pretty well, but I just wanted to add one thing: It's important to note that no experiment, currently running or planned for, expects to detect "gravitons", in the sense that the effects they expect to see are completely describable by classical GR — no need for quantum. Of course, it's possible that when LIGO or LISA start collecting data, they'll uncover some bizarre quantum effect at the length scales they're examining; but nobody really expects this to happen.
posted by Johnny Assay at 6:43 AM on March 28, 2006

QM and GR contradict each other (and they both work, annoyingly). Physicists are trying to bring them together. Looking for gravitons (and doing other experiments at the edge of GR and QM) has nothing to do with not "believing" in GR. It's just trying to start down the path of resolving the conflict.
posted by teece at 8:12 AM on March 28, 2006

A lot more physicists work daily with QM than work with GR.
posted by Ethereal Bligh at 4:22 AM on March 29, 2006

I'm not alone in my confusion (link is to current New Scientist letters page). Thanks to all those who replied, many of the replies remind me of my difficulty with Physics so many years ago. I remember my tutor explaining that Physics does not attemt to discover how the world works, rather it is a long (and rather boring) apprenticeship to be able to accurately predict how it will behave. That seemed to me a terrible retreat from the visions of Newton and Maxwell. A cop out, really. I remain hopeful that some discipline is endeavouring to understand and describe the underlying processes of the world, but it's clearly not modern Physics.
posted by grahamwell at 2:21 AM on April 6, 2006