Brian, this is a real point of interest for me. When I've gone to these training courses I always asked many questions. This topic was brought up many times and the way it was explained to me was that by sanding, your opening up the pores in the paint and when you apply more paint the solvents in the freshly applied paint permeate the sanded substrate bringing with them the active ingredients that form the chemical bond. I have taken the test panel and soaked it in reducer for 15 minutes. The sanded side is sticky and starting to lift on the edges...I then took a razor blade and around the edges and with a bit of effort I can remove the clear and the base. The Epoxy doesn't seem to be effected. The side that wasn't sanded, the edges are fine, the clear seemed a little stickier than before but a razor blade wouldn't touch it unless I gouged the clear. 10 minutes later, the clear wasn't sticky at all.
Now, I have been told the same story by different manufacturer's as to the benefit of sanding. They have all claimed that it isn't the size of the tooth that gives you the mechanical adhesion...it's more the number of teeth...that's why you get mechanical adhesion even with a fine grit of paper (a finer sanding footprint say with 1,000 grit than with 180 grit...thus more sand scratches and more teeth with 1,000). If chemical adhesion didn't occur by opening up a surface, would a paint company recommend using fine grits like 1,000 and even finer such as 1,500 on a clear blend on a sail panel? Also, I prep everything with 800 wet for base coat...some people swear that 400 grit gives better adhesion...I haven't had a peeler with 800 grit, and now that water born is out, PPG recommends sanding with 800 grit wet for base coat, why, enough of a grit to open the primer and because water born is so thin, 600 grit wet sand scratches have tendency to show...yet adhesion is not a problem.
Opening up primer by sanding allows the solvent to penetrate the primer, the solvents go through the primer, hit the metal and bounce back up. This also helps explain things such as solvent popping, poor adhesion of base coat if proper flash times aren't allowed and host of other scenario's that can effect paint.
I think we have all heard the term, paint needs to breath, the only way paint can breath is if it is porous to a degree. This is evident when a fresh paint job is waxed...the paint can't breath and eventually the clear or the color dies. Again, what we are doing by sanding is opening up the pores even further.
This how I've had it explained to me and it does make sense. One more note, I remember a prepper that did a color change on his car, every thing but under the hood, in the door jambs, inside the trunk and in behind the gas tank flap, he did however sand behind the flap. Within a year, the bottom lip on the inside of his gas tank flap didn't have any paint on it anymore.