In his book A Modern Introduction to Quantum Field Theory, Michele Maggiore affirms that
We require that all fields, independently of their transformation properties under the Lorentz group, behave as scalars under space-time translation.
I don't understand why we can assume so nonchalantly. Is there a way we can, instead, derive this statement in a mathematical fashion? I was thinking about the fact that the translation group is commutative, therefore every representation is completely reducible as a direct sum of one-dimensional representations... but I don't know if it is of any use.
I looked at other questions on the subject but they didn't clear my doubt: for example in this one the highest-rated answer says that the translation part of the Poincaré group acts trivially on the fields by $A(x)\mapsto A(x-a)$ but why is that so? Why can't we choose a transformation law $A(x)\mapsto A'(x-a)$ where $A'\neq A$?