I understand that asking pure math questions on the physics site is discouraged and I've seen many questions moved over to math. However, often times such questions are more suited for physicists to answer since they encounter those types of questions more often. Furthermore, they are more likely to answer the question using tools that a physicist would be familiar with. This is certainly the case for example is in this post and there are many other similar examples.

Should it really be disallowed to ask a math questions on the physics stackexchange if you would like a physicist's perspective on it?

Update: I just wanted to give an example for what my issue was. I recently needed some help with a math question from my QFT book, so I posted it here. However, the response I got was (as I expected) very formal, and not the approach a physics student would normally take.

• I agree, the language of physics is mathematics, and the more advanced the topic considered, the more math is needed to understand and talk about it at a technical level. So IMHO Physics SE is currently way too aggressively migrating technical questions involving mathematical concepts and ideas which are of paramount importance to understand advanced physics, away. In particular in theoretical physics, no clear separation can often reasonably be drawn, mathematical and physical ideas strongly interact and overlap. I would appreciate reviewers to refrain from too strictly separating things. – Dilaton Feb 12 '14 at 16:10
• @Dilaton that sounds more like an answer than a comment (and so it should be posted as such) – David Z Feb 12 '14 at 16:15
• @Dilaton "way too aggressively"? Have any proof? Have a look at the snapshot in my answer, I don't see any questions of the type you describe there. – Manishearth Feb 12 '14 at 16:18
• I never posted here (yet !) but a question I posted on Math.SE was never answered, the question would have been rejected at MO; and today I found the answer in a book in the math-physics section of my library. I wonder if this question would be rejected at Physics: math.stackexchange.com/questions/636065/… – Sergio Parreiras Feb 12 '14 at 17:56
• A relevant, if controversial, quote by the great V. I. Arnold: Mathematics is a part of physics. Physics is an experimental science, a part of natural science. Mathematics is the part of physics where experiments are cheap. – Logan M Feb 12 '14 at 19:30

I agree, the language of physics is mathematics, and the more advanced the topic considered, the more math is needed to understand and talk about it at a technical level. So IMHO Physics SE is currently way too aggressively voting to close for and migrating technical questions involving mathematical concepts and ideas which are of paramount importance to understand advanced physics, away (sometimes even against the judgment of established experts on the topic at hand)

In particular in theoretical physics, no clear separation can often reasonably be drawn, mathematical and physical ideas strongly interact and overlap. I would appreciate reviewers to refrain from too strictly separating things.

In addition, questions involving mathematical topics/concepts the "smell" too physicsy are often ignored on Math SE, as said for example here such that they find no answers if they have not already obtained some before being migrated, as the mathematicians there are not too interested in maths that is mostly useful for physicists. It also has to be considered that mathematicians often use different language/terminology, strategies to approach a problem (mathematicians consider physicists often to be sloppy, whereas the very rigorous point of view of mathematicians is often not directly useful for a physicist to solve his issue if interest), etc. ...

• (a) Evidence for your second sentence? Eg what you mean by a technical question that got migrated wrongly; that way we can notice and work on any issues in the migration system. (b) As far as the zeta function question goes, it's of the type that would never have been migrated away; it obviously has physics content.(c) When there is no clear separation we don't migrate, but many a times there are questions about mathematical proofs where the physics of the question is understood. These get migrated. – Manishearth Feb 12 '14 at 20:59
• One thing that can help reviewers 'refrain from too strictly separating things' is for 3k+ users to go to the review queue as often as they can to help deal with the tide of slush questions, and ideally not to just cast a single Leave Open vote. Reviewers are more likely to jump the gun on such questions if they've just slogged through ten mindless please-do-this-textbook-problem-for-me requests. – Emilio Pisanty Feb 13 '14 at 8:30

It depends on the question. Many a time, the physics of the question is already done, and the mathematics is jut regular math, which can be handled equally well (or better) by those on Math.SE.

There are those questions which are really physics-math and not math-math (eg some kinds of theoretical physics), and afaict they aren't migrated unless they have other issues.

Here's a snapshot of the recent migrated questions, to paint a better picture of the migrating practices:

(10ks can have a look at the page directly here)

The criterion for a question to be on topic here is that is is about physics. That's all. So yes, if a question is about math and not about physics, it shouldn't be allowed here.

Merely being more suited for a physicist to answer isn't really relevant; for one thing, if that were our criterion, it would allow many other kinds of questions that I think everyone would agree are off topic here. We've seen this happen in the early days of Stack Overflow, where people would ask questions that weren't about programming simply because they thought programmers would have the right perspective to answer. They were widely considered to be "polluting" the site.

Now, there is a fine line between advanced theoretical physics and pure math. To that end, there are certain topics, which one might consider purely mathematical, but which the community has effectively decided to include in the scope of physics, including renormalization and group theory. So sometimes it's a bit of a judgment call. But overall, the principle is that the question must be about physics to be on topic here.

• I am both a mathematician and a physicist and I do not think that the criterion "to be on topic here is that it is about physics" is so easy to apply. There are issues, in physics, but strongly entangled with mathematics that cannot be handled by mathematicians, because they simply cannot "understand" them and mathematicians cannot answer in a way a physicist consider appropriate. Almost all issue related with renormalization and regularization are of this nature. Especially in view of the language they are discussed by physicists. – Valter Moretti Feb 14 '14 at 7:48
• An example is this question math.stackexchange.com/q/672816 What, in practice I think it happens, is that only (mathematical-theoretical) physicists can answer with sensible answers. But just the fact that the question has migrated makes more difficult to answer it for several, even psychological reasons. I believe that (re)moving such questions from phys.SE is an error. – Valter Moretti Feb 14 '14 at 7:53
• @V.Moretti FWIW I think that that falls under the type of math questions that we consider to be physicsy enough, however it has a different problem: It's just asking for a step-through of the proof. Math.SE is more tolerant of such questions, we aren't. So it made more sense to migrate that than to just close it. Any more examples? – Manishearth Feb 14 '14 at 16:57
• @V.Moretti Renormalization and all are not migrated, unless they are off topic for other reasons. We consider math-that-only-physicists-can-do to be physics when it comes to migration. – Manishearth Feb 14 '14 at 16:58
• @ Manishearth You wrote: "it has a different problem: It's just asking for a step-through of the proof. " No, it was not. It could seem to a superficial reader. The point is that that question was and is very problematic from the pure mathematical viewpoint (here I am speaking as a mathematician). The identity written by the OP is false and to have some sense it needs an interpretation that only a physicist expert on regularization procedures can understand. For a mathematician that identity is a simple nonsense. – Valter Moretti Feb 14 '14 at 17:16
• For this reason I believe that moving that question was an error and math.SE is the wrong place for that, at least more inappropriate than physics.SE – Valter Moretti Feb 14 '14 at 17:17
• @V.Moretti but if the identity is false, how does it make any definition whether a physicist or a mathematician is looking at it? In any case, if in general you believe that being about physics should not be the criterion for a question to be on topic for this site, you should probably make a meta post requesting to get that changed. – David Z Feb 14 '14 at 17:58
• @David Z No kidding please :-), you know very well that physicists handle "mathematical" objects that do not exist in the sense the symbols seem to suggest. Feynman integral does not exist as an integral with respect to any measure. Similarly $Tr \log(-\Delta)$ does not exist since the operator is not trace class. Nevertheless that written identity has a precise meaning in physics and corresponds to a precise prescription, different from the one the used symbols seem to suggest. Here a physicist understands, a mathematician does not. I will consider your suggestion. Thanks. – Valter Moretti Feb 14 '14 at 18:20