Let $x$ is the notation of an element of argument domain.
Now $Ax$ and $Mx$ is predicate for the sentence that "$x$ is an American", and "x is a man", respectively.
I want to symbolize the sentence that "Some americans are men"
There are two ways. One is $\exists x (Ax \wedge Mx)$, the another logical syntax is $\exists x(Ax \rightarrow Mx)$.
In my textbook the only first syntax is right symbolization, second is not.
However, when we focus on the second syntax, it should be also true for an element $d \in D $, but $Ad$ is false.(vacously true), and I think it is well done.
Why does it happen?
Here's another way to think about it.
If you want to prove the statement $(\exists x)(Ax \to Mx)$, then you need to find an $x$ such that $Ax \to Mx$ is true. Well, if $Ax$ and $Mx$ are both false (for instance, $x$ is a woman from another country - say, Angela Merkel), then $Ax \to Mx$ is true.
But have you really proven that there is a man in America by pointing out that Angela Merkel is a German woman?
No. So something is wrong with your formulation.
Note, that as a rule, $(\forall x)$ usually has $\to$ in it, and $(\exists x)$ usually has an $\wedge$ in it.