We can find the opposite to be impossible i.e. show that assuming for ex. De Morgan's rule false contains a contradiction.
An example: Let's assume (AB)' in some case to be different than A'+B'. That means there's at least one pair A,B which make the following exclusive or ((AB)' xor (A'+B')) =1. We remember that X xor Y = X'Y+XY'.
Thus with some A,B there must be ((AB)(A'+B') + (AB)'(A'+B')')=1
The left half cannot be 1 because it's ABA'+ABB' which both are zero for every A,B.
Thus (AB)'(A'+B')' must be =1. That means (AB)' and (A'+B')' must both be =1.
(AB)'=1 means that AB=0, so either A or B or both are =0. Then A' or B' or both are =1.
In that case A'+B'=1, (A'+B')'=0. Just a couple of lines above this (A'+B')' was 1. That's the needed contradiction. So De Morgan's rule is true.
I guess the truth table is much more easier way.
Another way is to perform direct algebraic manipulation by catching 1 as a multiplier or zero as added, both not affecting the logic value. An example:
X'+Y= 1(X'+Y)=(X+X')(X'+Y)= XX'+XY+X'X'+X'Y= 0+XY+X'+X'Y= X'+X'Y+XY= X'(1+Y)+XY=X'+XY
We have proved that X'+XY = X'+Y. It's possible that one makes this calculation without anybody asking and really by accident founds a theorem, but as well someone can have asked a proof.