Consider two genetic terms in a generic Poincare invariant quantum field theory:
A trilinear term of the form $\phi_1\phi_2\phi_3$, and
a quartic term of the form $\phi_1\phi_2\phi_3\phi_4$
where $\phi_i'$s can (i) all be different, (i) some can be same and some are different or (iii) all may be different. Let us also not declare whether $\phi_i$'s are all scalar, vector or fermionic.
Is it correct to say that the first term always contributes only to decays and the second term always contributes only to scattering?