Changing the subject (phew!) ... my submission for the pedantry award:
Flight testing is done within an 'envelope'. Now't to do with envelopes, but the shape of the bits of the graph of load (y axis) against speed (x) that you have investigated is usually open-envelope shaped. The 'point' of the 'flap' being at the origin (no speed, no load).
So 'pushing the envelope', beloved of film makers, corporate bullshitters and folk who don't know anything but like to do buzzy talk, derives from flight testing where the engineers seek to push a single factor beyond the confines of the envelope - which is where they've already been and the aeroplane hasn't broken or bent.
So, maybe, a 2g turn at 600 knots goes to a 2g turn at 620knots. All done carefully and methodically, then back to base for the write-up and analysis before going to 640 knots.
The small increments are essential because things like flutter often have a sudden onset and can be completely destructive - loss of prototype!
I think it came to non-engineering prominence with Tom Wolfe's 'The Right Stuff' in the 1980s.