Ill go on a limb and say the NEC has never been concerned about fault clearing to that level in the past. I do not believe breaker disconnect times have ever been a concern... in fact we can legally have #14 on a 40 amp breaker feeding a motor.
As long as the OCPD clears before the wire reaches dangerous temps I don't see a fire taking place no matter the fault current magnitude.
What makes this a contrversial mystery is that it has never been proven whether or not faults taking there time to trip a breaker result in fire out in the real world...
The issue would no doubt be a rare occurrence, probably why it isn't directly addressed in the code (and IMO doesn't need to be). How common is an extremely long run, where conductors are _not_ oversized for VD? Surely there are many installations like this (but a very very very small percentage of all branch circuits in the country) - then how many of those installations have been subject to a fault - then how many of those resulted in damage warranting investigation - then in how many of those was the investigation carried out with time-to-clear in mind (or remaining evidence after fire even able to invoke the idea of too-long time-to-clear)??? Indeed, it would be hard to find a scenario where extended time-to-clear could be proven as the cause.
How high is that probability? This is a really interesting question, and I am not saying your wrong, but where in reality would a short circuit taking its time to clear a fault start a fire? What would that scenario look like? What materials would be involved?
Ever done any MIG or TIG welding? I can assure you, starting a fire with prolonged arcing is easy. As worst-case scenario, picture a cord-and-plug connected pendant-mounted HID fixture above dry animal bedding, cord damaged by age/rodent, sparks fly, big hot fire.
...In theory the NEC does address it (depending on how you look at it), but not in the way most expect.
This is sort of my point, once you start playing around with NEC ampacity tables you must consider more factors than just cable temp ratings or thermal buildup thresholds - among others, I think that fault current/time-to-clear factor is an important one and especially with the smaller conductors.
Noting that the NEC is "written in blood" I don't think time-to-clear on long runs really needs to be addressed, ampacity tables have a proven history, but if ampacities are increased under certain conditions then all factors must be considered. It's a huge can of worms IMHO. I don't see the value in increasing ampacities (or decreasing, for that matter) introduces a new world of risks.