This never ceases to puzzle me. Why, when we observe the behaviour of energy in more 'material' objects, should the rules change when the 'energy carrier' decreases in density? Objects made up of 'loosely' tied atoms absorb energy because of the volume of 'free electrons' they contain, and objects of densely packed atoms (namely crystalline structures) let light pass because they have few energy-absorbing free electrons; crystalline structures are in fact photon carriers.
When we observe the behaviour of light through crystalline structures, namely lenses, we can see that they can 'bend' light depending on their form. Light passing through lenses in fact not 'bent', but deflected: light exiting a crystalline structure will do so in a direction away from the thickest part of the lens, or the part of the lens containing the most speed-reducing atoms.
Doesn't the vacuum we call space behave in the same way? It is known that gravity can 'bend' light, a phenomena often attributed to the hypothetical gravitational qualities of photons themselves, but what if it was not the photon itself that was being deviated by gravity, but its carrier?
This makes perfect sense to me. Take, for example, light travelling past a black hole: if the gravitational pull is greater towards the centre of the black hole, so is the mass density; light has more density/gravity to go through on its side towards the black hole, so its path until the point where the gravitational pull is strongest, it will be deflected away; once past the gravitational apogee, if the black hole is perfectly spherical, the light will be deflected back towards its original path.
I wouldn't be surprised if gravity has no effect on light at all. If light depended on a 'carrier' that exists even in an environment we consider to be a 'vacuum', we could do away with the 'electromagnetic quality' theories about photons; photons would become a form of energy whose transmission depends on the quality of its carrier, and would behave just like any other energy known to us.
This leads me to believe that there is no such thing as 'nothing'. If the 'vacuum' of space was in fact a sea of inert 'perfect state' matter, or a material that some scientists are beginning to call 'dark matter', this would simplify the spectral map, and behaviour, of our universe's elements enormously.