Well, in my NEC (2002),
220.14 refers to motor loads, so I take it that you're referring to 220.3(B)(9):
Receptacle Outlets. Except as covered in 220.3(B)(10), receptacle outlets shall be computed at not less than 180 volt-amperes for each single or for each multiple receptacle on one yoke. A single piece of equipment consisting of a multiple receptacle comprised of four or more receptacles shall be computed at not less than 90 volt-amperes per receptacle.
Now, since this appears in Art. 220, I look at the scope, which says:
220.1 Scope. This article provides requirements for computing branch-circuit, feeder, and service loads.
This clearly says that Art. 220 covers
computing loads, or load calcs, as we call them. This refers to allowing adequate service ampacity to be able to supply a given amount of power for receptacle loads.
Nowhere do I see a requirement of how to wire circuits for general-use receptacles, or to limit the quantity of receptacles per circuit. This is a capacity allowance, not a wiring instruction.
Look at it this way: Let's suppose we wire a building with the minimum requirements in mind. Let's also suppose we work it out so the receptacle circuits happen to coincide with the 13-on-a-20a-circuit theory. Let's say we have five circuits with 13 receptacles each, for a total of 65 receptacles.
According to the 13-on-a-20 theory, if for some reason, one circuit stops at 12 receptacles, we would have to run a sixth circuit just for that last remaining circuit. Personally, I would likely not even reach 13 receptacles on one circuit in a non-residential installation, but that's not the point.
Now, if there really was such a physical limitation, I would imaginge it would be clearly spelled out, either in 210.7 or Art. 406. How many times have we argued that the NEC is not a design manual, and for those who complain when it seems to cross that line, that it should not be?