Right, i'll try and explain it a different way:
If you had a metal cased CU, and it was loaded to 50A, you'd have a current of 50A on each tail, well technically 50A on one tail and -50A (because the current flow is in opposite directions), now if each tail went in through a different K/O, the sum of currents through each K/O would be 50A, and this would act like a current transformer, with the tail being the primary, and the metal of the hole surrounding it being the secondary.
if both tails went though the same hold, the sum of the currents through that hole would be 50A + -50A =0A, as there is no net current on the 'primary' coils, there is nothing induced in the 'secondary'
If there was no option but to put the tails through separate holes, then you could cut a thin slot between the two holes, effectivly transforming it into one hole.
Thinking of it in terms of a clamp meter is probably better than a transformer, clamp one tail, and you'll get 50A, clamp both and you'll get 0A
if you draw out the two way switching circuit spoken about here, you can see that if the DI single went through a separate hole than the T&E, then like in the CU example that the currents through each hole would not be balenced, but with it being a lighting circuit and currents being small, the effect is likely to be miniscule