It (most likely is going to be) my field of expertise, too... nice to see someone else like this.
Liquids (and gases) are fluids.
You're right about the polarizability causing intermolecular forces (London dispersion forces in case of air), but those are the very-short-range forces I was referring to. If you compare mean free path (order of tens of milimeters) to approximate distances from which non-negligible polarization arises (order of angstroms), you'll find out that polarization really only governs molecules when they're "colliding".
Viscous forces arise from forces between molecules, it's not good to put these together with elemental forces like London's or G. And internal gravitational forces are even more negligible than London forces. Take 2 molecules of N2 for example, at their mean distance, and calculate the ratio between energy of their field vs. mean kinetic. I get about 10^-30. Energy is not the best for comparison, but there's no doubt in this case. This is caused by both bodies having very low masses. In comparison, the same ratio for an N2 molecule close to the Earth's surface and Earth, vs. mean kinetic energy is 10^4.
So no internal gravity, and close range London dispersion (which btw drops much faster with distance than typical Coulomb force). Diffusion is right, but if there was no gravity, it'd be diffusion air-vacuum, which fits your explanation with sand, too.
I don't dare to definitely say much about how sand acts in comparison (what I said about that before were just "educated guesses"
), but I'm sure about typical air.