Elevation is used to differentiate groupings of content using hierarchy and visual depth
In the real-world, nothing is perfectly flat. A page of paper on a desk has subtle shadow that indicates depth. We are attempting to model that in our interface. Layers live on different planes that casts a shadow that matches the depth of the layer.
We have five different levels of elevation, starting at Base which is the default level and grows in depth as the layer moves off the page.
If we model these on an axis, the height of each level becomes clearer.
// Shadows and Border.elevation-0 : Base.elevation-1 : Flat.elevation-2 : Raised.elevation-3 : Hovering.elevation-4 : Floating// Shadows Only.shadow-0 : Base, Flat.shadow-1 : Raised.shadow-2 : Hovering.shadow-3 : Floating