It's the same problem with trying to reconstruct dinosaurs, with probably the same solution in terms of public communication -- producing a _range_ of possible reconstructions based on the available evidence.

That said -- I think we actually do have more indirect evidence than what the reconstructions used -- in fact 3 separate lines of evidence A) paintings of statues B) contemporary descriptions of statues and C) contemporary paintings in general. All of which suggest that the coloring would have been more subtle and realistic.

I think if we had contemporary paintings of dinosaurs with feathers and contemporary accounts in writing that dinosaurs had feathers, but no feathers in the fossil record, you would still be fairly justified in saying that dinosaurs probably had feathers.