Graviton: Is It Possible That It Does Not Exist?

[relativityfan]

hi,

is it possible that the graviton does not exist?. except the success of quantum field theory, and the usefulness of graviton in string theory, are there other good reasons to hypothesize the graviton? thank you

in string theory, gravitons are different because they can travel outside our brane, right?

 

[bcrowell]

Historically, there was a long period during which Bohr and his followers wanted the atom to be quantized, but wanted the electromagnetic field to be classical. You get severe logical problems with this kind of theory. For example, say a classical wave of ultraviolet light with energy E impinges on a metal surface, and E is equal to 1.01 times the minimum energy needed in order to liberate one (quantized) electron from the surface (i.e., produce the photoelectric effect). By conservation of energy, the UV wave should only be able to eject one electron. But since the UV wave is purely classical, it can't carry the quantum-mechanical correlations that would tell electron A that it can't be ejected because atom B, far away, was the one that already got ejected. Therefore you get a violation of conservation of energy.

If you substitute a gravitational wave for the electromagnetic wave, I think you get exactly the same kind of contradictions if you assume that gravity is a purely classical field interacting with quantized matter.

 

[inflector]

For example, say a classical wave of ultraviolet light with energy E impinges on a metal surface, and E is equal to 1.01 times the minimum energy needed in order to liberate one (quantized) electron from the surface (i.e., produce the photoelectric effect). By conservation of energy, the UV wave should only be able to eject one electron. But since the UV wave is purely classical, it can't carry the quantum-mechanical correlations that would tell electron A that it can't be ejected because atom B, far away, was the one that already got ejected. Therefore you get a violation of conservation of energy. (emphasis mine)

I'm not following this analogy, especially the bolded sentence.

What are the quantum-mechanical correlations and electron A and B and how do they relate to the classical wave emitter and potentially liberated electron?

Is there any chance you can break this out a little more explicitly?

 

[PAllen]

See:

http://arxiv.org/abs/0803.3456

For a serious (speculative) consideration of this possibility.

 

[kexue]

In the linearized version of general relativity, when gravity is weak and metric and Einstein equations are examined only to first order, gravitational waves are predicted. Further, when a field of particles would be placed in these gravitational waves, the field of particles would oscillate in such way that is invariant after a rotation of 180 degrees. This suggest that the gravitational fields once quantized are described by spin-2 particles.

Neither gravitational waves, nor spin-2 gravitons have been observed by experiment yet, but are predicted by linearized general relativity.

But only in this linear approximation, gravitational waves and gravitations can give a precise meaning. (String theory assumes as linearized general relativity a flat background metric.)

 

[bcrowell]

But since the UV wave is purely classical, it can't carry the quantum-mechanical correlations that would tell electron A that it can't be ejected because atom B, far away, was the one that already got ejected.

What are the quantum-mechanical correlations and electron A and B and how do they relate to the classical wave emitter and potentially liberated electron?

Is there any chance you can break this out a little more explicitly?

I don't like the Copenhagen interpretation, but this is probably easiest to explain in the Copenhagen picture. In the Copenhagen picture, with both electrons and light quantized, what happens is that initially the position of the photon has not been measured. But then when the photon is absorbed by electron B, the position of the photon has been measured, and its wavefunction collapses, so that it has zero probability of being at the location of electron A. This is how you preserve conservation of energy (in the Copenhagen interpretation) -- the photon can't be absorbed by both A and B.

If you don't quantize light, then you can't have wavefunction collapse, and you don't have any mechanism for preserving conservation of energy. Electron A is quantized, so it's a matter of probability whether it gets ejected or not. Likewise for electron B. Since there is no mechanism to provide a correlation between them, the probability that both are ionized is simply the product of the two probabilities, which is nonzero. So with nonzero probability, you violate conservation of energy.

If light is quantized, then there is an anticorrelation between A's ejection and B's ejection. (In the Copenhagen picture, this anticorrelation is due to wavefunction collapse.) Because of this anticorrelation, the probabilities don't just multiply. The probability that both A and B will absorb the light is zero.

 

[PAllen]

If light is quantized, then there is an anticorrelation between A's ejection and B's ejection. (In the Copenhagen picture, this anticorrelation is due to wavefunction collapse.) Because of this anticorrelation, the probabilities don't just multiply. The probability that both A and B will absorb the light is zero.

And this relates to the issue of hidden variable theories and Bell's inequality. A common conclusion is that any attempt at a hidden variable theory must violate locality and/or causality. However, a while back I caught some papers that I can't find now, that argued you could equally have hidden variable theory that satisfied local materialism if you admitted negative propabilities and corresponding changes in probability laws. Whether this helps anyone trying to pursue such an approach ...

 

[inflector]

If you don't quantize light, then you can't have wavefunction collapse, and you don't have any mechanism for preserving conservation of energy. Electron A is quantized, so it's a matter of probability whether it gets ejected or not. Likewise for electron B. Since there is no mechanism to provide a correlation between them, the probability that both are ionized is simply the product of the two probabilities, which is nonzero. So with nonzero probability, you violate conservation of energy.

If light is quantized, then there is an anticorrelation between A's ejection and B's ejection. (In the Copenhagen picture, this anticorrelation is due to wavefunction collapse.) Because of this anticorrelation, the probabilities don't just multiply. The probability that both A and B will absorb the light is zero.

Okay, I see what you are saying. That makes sense.

 

[bcrowell]

Thanks, PAllen, for the link to the Carlip paper -- very interesting! I think the argument I gave in #2 and #5 is of the same flavor as the ones Carlip presents near the beginning. My reading of the paper is that he admits that you can't couple classical and quantized fields in a logically self-consistent way, but he proposes that our understanding of "quantized" can be modified so as to allow such a coupling. This seems reasonable. After all, we know that GR and quantum mechanics are logically incompatible, so it's clear that we need to abandon some fundamental assumption of either GR or quantum mechanics.

 

[PAllen]

Here is another paper, (which cites the Carlip paper and acknowledges correspondence) carrying further the argument about the possibility gravity is not quantized. If I understand parts of this (small parts), they propose alternative GR-quantum couplings that don't require quantum non-linearity (which was the approach taken in the Carlip paper). http://arxiv.org/abs/0802.1978

 

[czes]

In the linearized version of general relativity, when gravity is weak and metric and Einstein equations are examined only to first order, gravitational waves are predicted. Further, when a field of particles would be placed in these gravitational waves, the field of particles would oscillate in such way that is invariant after a rotation of 180 degrees. This suggest that the gravitational fields once quantized are described by spin-2 particles.

Neither gravitational waves, nor spin-2 gravitons have been observed by experiment yet, but are predicted by linearized general relativity.

But only in this linear approximation, gravitational waves and gravitations can give a precise meaning. (String theory assumes as linearized general relativity a flat background metric.)

Due to Verlinde's gravity as an entropic force there isn't a graviton and gravitational field is not quantized.
Why ?
I think, Verlinde uses thermodynamics in order to explain gravity and he doesn't wrote the gravitational field isn't quantized.
If I good understand Verlinde's idea the emergent space is created of the microstates which behaves thermodynamically. The graviton is not a fundamental particle but emerges as a change in a density of the microstates.
Do I overlooked something ?

 

[Finbar]

In the linearized version of general relativity, when gravity is weak and metric and Einstein equations are examined only to first order, gravitational waves are predicted. Further, when a field of particles would be placed in these gravitational waves, the field of particles would oscillate in such way that is invariant after a rotation of 180 degrees. This suggest that the gravitational fields once quantized are described by spin-2 particles.

Neither gravitational waves, nor spin-2 gravitons have been observed by experiment yet, but are predicted by linearized general relativity.

But only in this linear approximation, gravitational waves and gravitations can give a precise meaning. (String theory assumes as linearized general relativity a flat background metric.)

Furthermore linearized gravity may be quantized within the frame work of effective field theory. Meaning that we do have a perfectly predictive and consistent theory of the graviton. However when gravity becomes strong the effective theory of the graviton must breakdown i.e. once we cannot treat gravity as a small perturbation over some fixed metric.

 

[relativityfan]

After all, we know that GR and quantum mechanics are logically incompatible, so it's clear that we need to abandon some fundamental assumption of either GR or quantum mechanics.

logically incompatible? they do not match yet but i do not see why they are LOGICALLY incompatible, why is this?

 

[inflector]

Rovelli has some choice words from the first chapter of his book on Quantum Gravity:

QM and GR have destroyed the coherent picture of the world provided by prerelativistic classical physics: each was formulated in terms of assumptions contradicted by the other theory. QM was formulated using an external time variable (the t of the Schrödinger equation) or a fixed, nondynamical background spacetime (the spacetime on which quantum field theory is defined). But this external time variable and this fixed background are incompatible with GR. In turn, GR was formulated in terms of Riemannian geometry, assuming that the metric is a smooth and deterministic dynamical field. But QM requires that any dynamical field is quantized: at small scales it manifests itself in discrete quanta and is governed by probabilistic laws.

The entire field of Quantum Gravity is an attempt to reconcile these incompatible foundations.