$P(\text{student}, \text{bilingual}) \le P(\text{student})$ and $P(\text{student}, \text{bilingual}) \le P(\text{bilingual})$.
This is a consequence of the fact that $A\cap B\subseteq A$ and $A\cap B\subseteq B$ and for $A\subseteq B$ one has $P(A)\le P(B)$.
Depending on your assumption on the sampling population, if you have at least a student who is not bilingual and and at least a bilingual who is not student, then the strict inequalities
$P(\text{student}, \text{bilingual}) < P(\text{student})$ and
$P(\text{student}, \text{bilingual}) < P(\text{bilingual})$ also hold.
Since your problem does not imply any restriction on the sample space, I am inclined to say c). Which means that if you just go out and ask somebody on the street, you will find bilinguals who are not students and students who are not bilinguals.