Bill, As long as the distances aren't astronomical, and the currents aren't extreme, I can't see any engineering principles involved here. Each reverse loop or wye having its own AR is strictly for practical reasons: because two trains might be at different thresholds simultaneously, setting up an impossible requirement for both phases simultaneously from a single output.
I actually violated that rule on my own railroad. I have a double track reverse loop. Immediately adjacent to each end, turnouts join the two reverse loops to single track main line segments. The boundaries to the reversing loops are at the diverging ends of the turnouts. Each of those reversing tracks is about 27 feet long, able to hold three of my typical length HO gauge trains. Connecting them to one AR menas I can't have a train exiting westbound from either track while another is entering eastbound at the other end. I accepted that limitation, since I usually operate the railroad by myself, and never felt the need to have trains entering and leaving those tracks at the same time.
BTW: the same requirement exists for each reversing segment, be it a loop or wye. Trains can't cross both boundaries at the same time, at least not powered units, or lighted or other cars with multiple metal wheels connecting to the power from each rail. If trains connect the rails across both boundaries at the same time, a short is inevitable. Usually, it's enough to have to observe the rule for one reversing loop or wye, without adding the needs of another to the equation.
So, the core reason is, having separate ARs for each reversing section prevents a lot of complex restrictions when running multiple trains.