First, you say you study math, but you are asking about science. They are not the same. In math we seek proof, but in science it is evidence that we are after.
One consistent view of math is that it is a completely human creation and works with well defined assumptions (axioms) and methods of proof (logic). That makes it a closed system, where "proof" means provably consistent with the axioms using agreed upon proof rules.
Science is very different, since it attempts to answer questions about the observable universe, which doesn't have an agreed upon definition. It just is. We observe gravity. We attempt to measure it. We make hypotheses about it and search for evidence that our hypothesis is correct. But that isn't proof. Einstein comes along and changes the game, with new hypotheses and new evidence which is "hopefully" more accurate than that which Newton supplied.
At one time the universe was (thought to be) composed of atoms (indivisible units). Now we have quarks and such. How do we explain those. We can't actually prove our hypotheses, but we can supply evidence of various sorts. But a scientist will always admit the possibility that someday someone will supply a more accurate (closely measurable) understanding of "stuff".
But that isn't like the Pythagorean Theorem, which, once proved, stands for all time, unless we change the underlying axioms, in which case we are studying something fundamentally different.
There are no questions that are "off limits" in either math or science. But there are questions that are more valuable when asked and proof (math) or evidence (science) is provided.
The valuable questions are those that provide insight into their domains. Special Relativity did that. The Derivative did that. They help us ask the next question in both math and science.
To build a math or scientific career, however, you need to value "answerable" questions so that you don't get stuck. For that, a novice, perhaps an undergraduate, normally needs a guide to lead them to the edge of what is "known" in either field. But the meaning of "known" also differs. In math, it is something provable from the axioms. In science it is something with sufficient evidence that it becomes generally accepted (for the moment) by experts. Note that Special Relativity wasn't immediately accepted in the scientific world. It took time to gather the evidence.
In both math and science we seek a kind of "truth" though it is different in the two domains and approached in different ways.
Journals seek to publish novel work. That is, work that is new in some way and beyond the ordinary. They have no interest in publishing proofs of student exercises, for example, as the techniques are well known and the results unlikely to provide new insights to experts. Similar, in physics, a new theory might seem "interesting" and the methodology of providing evidence might provide insight into other problems of the day. In science, even something new but as yet unexplainable, might be interesting enough for publication. If something unexpected pops up at CERN, for example, folks want to know about it even before a coherent explanation is found.
