Would a model have to have bones in its face in order to be faceposed?
No. Source faceposing works via morph targets and not skeletal animation.
What exactly are “morph targets”?
Basically it just finds the differences between vertex locations and moves them accordingly.
The vertices that are in control of the face are the morph targets, I believe. They’re called Shape Keys, and you can read a little bit more at the Valve Developer Wiki. This page has the overview on making characters (models) in XSI Mod Tool, which includes information of what you may be seeking. If you’re making them in a different program, like Blender, there should be a tab that has Shape Keys (should be in the object data tab). I don’t know about 3ds MAX, though.
Ah, beat me to it, ApocHedgie.