Wouldn't we notice by now? GPS would be affected if the speed of light changed. All kinds of speed-critical connections would be affected. Heck, even CPU speeds would change.
Read the thing. Speed slowed until it reached a lower bound, and has been stable since then. That's the theory described.
The TL;DR is "the cosmic background noise is too uniform; if light moved FAST in the early universe that could explain the uniformity; there have been theories; here is one."
tl;dr: Light and gravity both move at "c" locally and near us under this proposal, so we would notice no difference from General Relativity (GR) or local tests of Special Relativity. In fact, it's only the generality of GR that lets one call this a variable speed of light (VSL) theory rather than a time-varying speed of gravitational wave theory (see my last paragraph before the footnote below). The relevant time is confined to the very early universe.
Afshordi and Magueijo's argument is essentially a bimetric theory of gravity.
In General Relativity, there is a single metric and universal coupling of all matter, radiation, and gravitons (more on those in four paragraphs) to it. There is a tangent space (illustration at [0]) at each point in spacetime, and the Minkowski metric is induced on each of them, which is how we get Special Relativity locally. In the absence of any matter, radiation or gravitons, all tangent spaces are aligned with one another, and they all have universal extent; we call that Minkowski spacetime. Once you add in matter, radiation or gravitons, the tangent spaces contract and become unaligned, and you can no longer use the Minkowski metric to describe things outside a chosen tangent space -- another metric applies instead, and all matter, radiation and energy couple universally to that.
The crucial point though is that there is one metric, and the change of vector magnitude and direction from point to point in spacetime is described by it and applies to everything in the spacetime.
In a bimetric theory some matter, radiation or gravitons couple to one metric, and some to the other. In the authors' argument, gravitons couple to the second metric while everything else couples to the first. Moreover, the graviton metric only applies in the very very early universe and quickly decays into indistinguishability from the matter-energy metric. So General Relativity still applies everywhere that our tests suggest it's correct.
Bimetric theories have been studied since Rosen (of Einstein-Rosen bridges) proposed the idea in 1940, but they are pretty strongly disfavoured by evidence supporting the single metric of GR, most recently by the LIGO observations. Consequently most viable bimetric theories have the second metric operate in inaccessible regions like the very early universe. Afshordi & Magueijo's metric decays so early as to produce almost no observables at all, but atypically for modern proponents of bimetric theories, they deliberately predict some observables in the relic fields like the cosmic microwave background.
Crucially, the second metric in Afshordi & Magueijo's approach applies to gravitons only.
Gravitons are a feature of canonical general relativity and other perturbative approaches. Such approaches are perfectly fine except very close to physical singularities where (on a Feynman diagram) multiple loops of gravitons start appearing. In other words, you can quantize General Relativity's clasical spin 2 symmetry gravitational waves in an effective field theory (in the Ken Wilson sense), and in so doing reproduce General Relativity's results exactly as long as you aren't in strong gravity.
Strong gravity is not a consideration in A & M's proposal. However, unlike canonical general relativity's massless graviton, theirs has a mass in the early universe which decays to zero in the later universe.
In their proposal, in the very early universe, the tangent spaces at each point on a photon's worldline are different from those on a graviton's worldline. This means that for a graviton and a photon at the same point in spacetime, the magnitude and direction of the vectors in the tangent spaces differ, and pretty wildly. The "next step" for the massless photon can have a much (much, much) longer spacelike component than the massive graviton; the photon can exit the graviton's causal cone (however the mass of the graviton is s...
5 comments
[ 2.5 ms ] story [ 25.0 ms ] threadThe TL;DR is "the cosmic background noise is too uniform; if light moved FAST in the early universe that could explain the uniformity; there have been theories; here is one."
Afshordi and Magueijo's argument is essentially a bimetric theory of gravity.
In General Relativity, there is a single metric and universal coupling of all matter, radiation, and gravitons (more on those in four paragraphs) to it. There is a tangent space (illustration at [0]) at each point in spacetime, and the Minkowski metric is induced on each of them, which is how we get Special Relativity locally. In the absence of any matter, radiation or gravitons, all tangent spaces are aligned with one another, and they all have universal extent; we call that Minkowski spacetime. Once you add in matter, radiation or gravitons, the tangent spaces contract and become unaligned, and you can no longer use the Minkowski metric to describe things outside a chosen tangent space -- another metric applies instead, and all matter, radiation and energy couple universally to that.
The crucial point though is that there is one metric, and the change of vector magnitude and direction from point to point in spacetime is described by it and applies to everything in the spacetime.
In a bimetric theory some matter, radiation or gravitons couple to one metric, and some to the other. In the authors' argument, gravitons couple to the second metric while everything else couples to the first. Moreover, the graviton metric only applies in the very very early universe and quickly decays into indistinguishability from the matter-energy metric. So General Relativity still applies everywhere that our tests suggest it's correct.
Bimetric theories have been studied since Rosen (of Einstein-Rosen bridges) proposed the idea in 1940, but they are pretty strongly disfavoured by evidence supporting the single metric of GR, most recently by the LIGO observations. Consequently most viable bimetric theories have the second metric operate in inaccessible regions like the very early universe. Afshordi & Magueijo's metric decays so early as to produce almost no observables at all, but atypically for modern proponents of bimetric theories, they deliberately predict some observables in the relic fields like the cosmic microwave background.
Crucially, the second metric in Afshordi & Magueijo's approach applies to gravitons only.
Gravitons are a feature of canonical general relativity and other perturbative approaches. Such approaches are perfectly fine except very close to physical singularities where (on a Feynman diagram) multiple loops of gravitons start appearing. In other words, you can quantize General Relativity's clasical spin 2 symmetry gravitational waves in an effective field theory (in the Ken Wilson sense), and in so doing reproduce General Relativity's results exactly as long as you aren't in strong gravity.
Strong gravity is not a consideration in A & M's proposal. However, unlike canonical general relativity's massless graviton, theirs has a mass in the early universe which decays to zero in the later universe.
In their proposal, in the very early universe, the tangent spaces at each point on a photon's worldline are different from those on a graviton's worldline. This means that for a graviton and a photon at the same point in spacetime, the magnitude and direction of the vectors in the tangent spaces differ, and pretty wildly. The "next step" for the massless photon can have a much (much, much) longer spacelike component than the massive graviton; the photon can exit the graviton's causal cone (however the mass of the graviton is s...
Seems as unlikely as inflation theory to me.