Everything in the real world. Does your subject land anywhere on something other than a totally predictable straight-line part of the curve. Of course it does. You've got gradation curves inherent to
shadows and highlights, and what you do to midtone expansion affects your choice of how this works.
ASA is really arbitrary in this respect, even from one shot to another with the same film. What counts is how you actually distribute specific parts of the scene onto the curve geometry. The Zone System is
just a generalized way to classify how you go about this, which you can make as simple or as complicated as you with. Walking around with a calculator and some other model might work perfectly
well for someone else. But in effect, you are still altering your ASA whenever you change your endpoints
or midpoint. It's all relative. Hypothetically, I might even want to blank out the shadows ala Brett Weston, but gotta know what's going on at the bottom of the curve, regardless of ASA. One film will
do it, another will just leave you with mud, even though under ordinary circumstances you would assign
them exactly the same ASA. I guess you could go around with a pocket computer programmed with
calculus and fiddle with all kinds of silly integers or micro-hedgehogs in it or whatever. If that's the kind
of thing you enjoy, fine. I do plenty of densitometer work in the lab when it's actually relevant. But in
the field I want to instinctively know what's going to happen under a wide and often suddenly-changing
scope of lighting. The way I conceive of ASA might shift just as fast.