I can no longer resist going on the rant that electrofelon resisted.
I can't quite figure out where this new rule originally came from because it comes from a Correlating Committee note that refers to a first revision I can't find. That is, this language was originally proposed in different sections and ended up in 230.46 in the Second Draft. It may have originated with an input from none other than Mike Holt to move an existing requirement from 314 because it really belonged in 230. All this is to say, while I'm not 100% certain, as far as I can tell...
The original language had nothing to do with "pressure connectors for splices and taps" and was limited to power distribution blocks. The CMP decided to go above and beyond any public input or previously existing requirement and add the requirement for pressure connectors.
There's a reason that the original requirement only applied to power distribution blocks, as Robert Osborne from UL explained in a second revision comment. Power distribution blocks are fixed to the enclosure and thus need to be strong enough to keep the conductor from blowing out of them (due to magnetic forces) in a short circuit situation. But pressure connectors can bounce around with the wires, so "a conductor is not likely to pull out of the connector". Because the connector just moves with the wire.
Nonetheless, the CMP ignored Osborne's informed comment, placed an uncalled-for mandate on UL to do more research and develop a new standard, even though, as member CMP Paul D. Barnhart stated in a comment on his affirmative vote, "Free floating splices have been commonplace on the line side of the service disconnect for decades and the CMP does NOT provide evidence that this longstanding practice has been problematic."
(emphasis added)
The whole thing confirms that the code making process is based more on speculation than evidence or responding to real-world incidents.