Well the name refers to the relative incident and internal angles because that's what scientists first measured. The light slows down due to interactions with the electrons (mostly) in the material. It causes them to move which drags the light. You can model this as an interference from the light produced by the resulting electron movement. I don't see that as bouncing, especially not like the bouncing on the internal surfaces of fiber optics. But obviously it's not like anything we can tangibly understand, so whatever mental model works for you is cool.