To give some basis for a debate, this is mostly a copy of my answer to the corresponding question on Area 51. As I yet failed to follow this site as much as I planned to do, this may not be up to date with some aspects already established on this Meta.
TL;DR: Do not define history by age, but by what it is about.
First of all, while I mainly write about science here, this is to keep the text readable. Very similar (and at times stronger) arguments apply to math.
Science idealistically always aspires to be up to date and thus it does not need to care about its history. Or with other words: Science cares about how our current view of the world is and how it is backed up¹, but not about how it came to be. Relatedly, the other scientific Stack Exchanges mainly have questions about our current state of knowledge, current definitions and notations or, in the case of Academia, about the current state of the scientific community.
In contrast to knowledge, society – the main subject of regular history – cares a lot about its own past. Hence it makes some sense that recentness is not the same to this site than to regular history or even to define history by something other than age.
Thus I would define history for our purposes as everything that is not about the current state of knowledge, notations, definitions, the scientific community, etc., but about how it evolved or how it was. As another criterion, the best answer should only be able to change due to new insights on past events and should not even be conceivable to change due to new insights on how reality works (i.e., scientific progress) or new notations and definitions coming into use.
To give an additional criterion for questions like (heavily exaggerating)
How did the theory of X change during the last week?
one could require that there should be some reason to assume a relevant change of the state of knowledge in the respective time frame. However, I do not expect such questions to pose a big problem and I think that they wouldn’t be asked like this in the first place, if there were no such reason.
Why do we assume X?
When and why was X first assumed?
is on-topic. However, for the latter, X should not be such an entirely brand new concept that there is no doubt that the answers to both questions are the same. In a rapidly evolving field, a few years might suffice for this, though.
What is the prevalent theory for X?
What was the prevalent theory for X, Y years ago?
is on-topic. Again, there should be a reason to assume that there was a significant change in the last Y years.
Making graphene was as easy as removing layers with scotch tape and it's structure was postulated for years, why was it so difficult to discover?
is clearly on-topic by the above definition as it is about the past state of knowledge.
Why was Pluto demoted from the list of planets?
is on-topic, though arguably not such an interesting question (but that’s for votes to decide and not a reason for closure).
¹ and of course how it can be improved