Talk:Chapter 7/@comment-76.185.171.168-20201008122913/@comment-37.11.55.91-20201009070218

This is the idea I have:

You consider the same formal system, only with an extra axiom P. If from that you can derive Q, then from the initial system you can derive P=>Q. The derivation of Q from P remains only a "fantasy"; its steps are not theorems of the original system.

It could also be thought as a complex typographical rule, as question 3 suggests.