41

I study mathematics as my major subject and theoretical physics and statistics as my minor subjects. I found that, sometimes in physics or statistics lectures, the lecturer makes mistakes, like forgetting to prove that a series converges, or computing multi-dimensional integrals by using only one path. Once I spent three weeks to find a correct reasoning why one particular series converges. Should I say anything about these mistakes to the lecturer?

  • 13
    And then you hear "we see this series diverges to infinity. Let's sum it". And in some situations, this is a perfectly reasonable thing to do, since you sum it in some sense that has physical interpretation. – Davidmh Sep 23 '15 at 9:06
  • 54
    "forgetting to prove" something mathematical is not a "mistake" in physics. The only "proof" I am interested in as a physicist is experimental validation. – Calchas Sep 23 '15 at 13:37
  • 36
    Note that for every student who wonders about the convergence of the series, there is a student who is completely baffled by all the math, and struggling to cope. This student needs a high level, intuitive overview, which is what the unrigorous lecturer is likely trying to provide. He's only got two hours. – Peter Sep 23 '15 at 14:33
  • 11
    Particularly in engineering, it's common to learn these concepts and their proofs rigorously in the math class. The physics lecturer then assumes that this knowledge is already well established. Often a physics course will have a math prerequisite for this reason. – user21268 Sep 23 '15 at 16:12
  • 36
    "Once I spent three weeks to find a correct reasoning why one particular series converges." doesn't this answer it already? Are you studying convergence of series or physics? – Bakuriu Sep 23 '15 at 18:32
102

In physics we deliberately do not prove the series converges, because we are not interested in teaching concepts like convergence. Physics courses are not intended to be mathematically rigorous. It just is not one of the goals.

  • 94
    You may have misunderstood. We use series because they get the right result. And we do care that they converge, we just don't futz around with the issue during class. We learned about series convergence in math class, and the serious theorists worry about it when writing papers (experimenters like myself generally don't have to because someone else already did), but class time is far to precious to waste on the matter unless the bounds of convergence are going to be important in this class of problems. – dmckee Sep 23 '15 at 0:03
  • 28
    This is the standard answer and I upvoted it. However I have always been profoundly frustrated by this and felt locked out of most of physics because of it. Ever since high school, things were fully explained in my mathematics courses and very erratically explained in my science courses. We care about convergence of series because divergent series may be absolutely meaningless. How can understanding why the procedures one is performing are meaningful not be part of science education? – Pete L. Clark Sep 23 '15 at 1:16
  • 32
    It is especially frustrating because it seems to be, as is hinted at in the comments, an issue of science pedagogy rather than science itself. If really no one knows that a series converges, that's a huge problem: e.g. a problem of this kind got Feynman, Schwinger and Tomonaga their Nobel Prize. All leading physicists I've met understand convergence of series just as well as I do; the difference is that this knowledge is kept mostly private in physics pedagogy. I don't think physics places any less of a premium on "true understanding" than mathematics, so this is a weird state of affairs. – Pete L. Clark Sep 23 '15 at 1:19
  • 14
    @PeteL.Clark And the physics department standard reply is that the math department is supposed to provide all those tools. The students can (and should) check that the claims made by their physics profs do indeed hold, but on their own time. (There's already not enough lecture time to bring undergrads past the year 1920 or so without the extra math.) Otherwise what good are all those math classes the physics majors are required to take? If the math department doesn't think this is their responsibility (as I've seen and is fair enough), then the blame falls on interdepartmental communication. – user4512 Sep 23 '15 at 6:35
  • 22
    I didn't say this material should be covered during lecture time (and I think the OP's characterization of this as a "mistake" by his lecturer is inaccurate; by the standards of the respectable, intelligent community of physics educators, this is certainly not a mistake). Probably it should not be, most of the time. What frustrates me is that this material -- well known to many physicists -- is so thoroughly absent from physics textbooks. The knowledge is there but it is not being conveyed. This is a shame, and it turns off many mathematically minded people...I think unnecessarily. – Pete L. Clark Sep 23 '15 at 7:02
74

Let me add a thought following the other post, which asserts that this is not a mistake, pedagogically. Take this as a given: you don't need to "call out" the professor for failing to teach "properly." It is, however, still the case that you personally are wanting to dig into the mathematical foundations of these concepts more deeply, and find it important to your comprehension.

That's great! You might learn something really interesting, and might set yourself on a path to become a person who makes scientific advances by attacking these sorts of questions.

Now, I would suggest approaching your professor from that perspective, instead of considering it a problem with their teaching. Ask if there are books or other resources that the professor would suggest where you can learn more about the proofs behind these assertions. If the professor doesn't have good suggestions for you, try looking in places like Physics.SE. If you can't find a satisfactorily rigorous proof, it may well be that it does not exist (unlikely, but it happens), and that may be an interesting opportunity!

  • 1
    For lists of physics books geared toward mathematicians there is this question on Math.SE (advertisement: where one can even find an answer of mine :-) ). – Massimo Ortolano Sep 24 '15 at 5:41
30

Echoing parts of the other answers, and some of the comments: first, it is inaccurate to declare omissions such as "proof of convergence" a "mistake". There simply is no absolute obligation to verify that all parts of the mathematics work as a physicist expects for other reasons. Yes, you or I and others might want to see the proof, that is, mathematical causality, but this is simply not obligatory. (Conversely, we can prove things without direct physical manifestations or physical reasoning...)

In fact, "convergence" is merely a simple form of what one might want, and itself not obligatory (much less its proof). Indeed, I have read that Poincare discovered in the late 19th century that a series expansion of a solution to a differential equation used for many decades (successfully) in celestial mechanics did not converge. Not that its converge was difficult to prove, but that it definitely diverged. But/and people had been getting correct numerical outcomes. Well, it was an "asymptotic expansion", ... but/and such expansions are more delicate in some regards (e.g., term-wise differentiation) than convergent power series, and the mathematical details were not filled in for several decades.

Another example is P.G.M. Dirac's book on quantum mechanics, which used distributions and unbounded operators in manners that would not be justified for 20 years (in the work of L. Schwartz). I have read that J. von Neumann and others were considerably disturbed by the lack of "rigor", or even the pretense of it, which motivated them to try to provide such... Nevertheless, the predictive and explanatory power of Dirac's work was unquestionable, and it would have been ridiculous to have dismissed it because he couldn't provide proofs, or didn't care to.

As remarked above, it really does appear to be that hard-to-justify mathematics is fairly tolerable when it quasi-magically predicts physical details, or quasi-magically proves to be an accurate book-keeping or computational device for observable physical phenomena.

Yes, we should think very differently when/if we aim to "subvert" such mathematics to purely mathematical situations, where there may be no genuine physical phenomenon to observe and test. No, I do not have that physics-y intuition that suggests (to my perception) outrageous mathematical manipulations, so I myself definitely need either or both pithy examples and persuasive (!) proofs that assure me there's some "causality" beyond the literal tangible world. But, in fact, history suggests that much interesting mathematics has come from "outrageous" mathematical stunts by imaginative physicists, so such stuff is a good source!

And, yes, sometimes the purely mathematical justification for obviously-necessary mathematical tricks in physics is far more sophisticated than the immediate physical explanation/motivation/phenomenon. Sure, sometimes the mathematics is not hard, and simply omitted due to lack of interest. Sometimes the mathematics is profoundly difficult, or in fact impossible in a particular year with technical limitations of the time. That fact, that has appeared over and over, is philosophically and scientifically provocative in itself, in my opinion.

So, yes, I, too, have been disturbed by reading physics-y accounts that did (to my perception) crazy mathematical things. Long ago, I thought that this was a definite failing, and that rigor was required, and possible. By now I see that these situations are much more complicated than that, and that gauging any particular instance may be unexpectedly non-trivial!

  • 4
    paul: This is a thoughtful answer. First of all, you're right that divergent series are much more useful in physics than they are in most branches of mathematics. To my mind, this is an argument that physicists need to pay more attention to these issues than most mathematicians. It is true that asymptotic expansions can be used successfully even when the mathematical formalism is not known. It is also true that they can be used unsuccessfully, and there are examples of both... – Pete L. Clark Sep 23 '15 at 22:29
  • 3
    ...You're also right that sometimes the "rigorous" (not my favorite word...) mathematical justification of something which is being used by physicists is very sophisticated, that sometimes no one knows how to mathematically justify what is being physically used, and that this is not necessarily a bad thing: in fact quasi-magical physical predictions are one of the payoffs of cultural relations between mathematics and physics. I would like to see more that, not less. What I don't believe is that "it doesn't matter" whether physical theories have a solid mathematical grounding. – Pete L. Clark Sep 23 '15 at 22:33
  • 2
    ....Isn't that idea some kind of clubhouse handshake that older undergraduate physics students teach younger undergraduate physics students, a certain kind of machismo? And now some people here are repeating this. But of course it matters: brilliant people, some called "mathematicians" and some called "physicists" have worked very hard to put physical theories on a firm mathematical grounding. And yet there is little to no trace of these efforts in undergraduate physics. That's what I object to: not everything is known, but some things are. Those that are should be told. – Pete L. Clark Sep 23 '15 at 22:37
  • 1
    @PeteL.Clark, yes, of course, incidentally, there is some exclusionary riff (human sociology) that confounds everything. E.g., rationalizing limitations or failings as virtue. :) Undergrad and graduate-student perceptions of things (in my experience) tend to drift into caricature/over-simplification/machismo... if only because they're adolescent (nevermind the historical/evolutionary sense of adulthood at age 13, etc...) So, we have bowdlerization + clubhouse/insider-trading denial. – paul garrett Sep 23 '15 at 23:15
  • 1
    @PeteL.Clark Yes, please, please give us struggling mathematicians an answer (it is a blessing for my tormented soul to see that other mathematicians also have problems with physics). And please see also my comments on the other answer.... – resu Sep 26 '15 at 13:56
13

Without detracting from the merit of the more philosophical answers, here's a simpler practical suggestion:

  • Go see your professor after class or in reception hours.
  • Tell him that since you're a Mathematics minor, you find the mathematical reasoning important to follow.
  • Tell him that you sometimes cannot tell whether a step he makes is actually trivial, or might take a lot of time/effort to justify rigorously.
  • Ask him that, when he is making a 'mathematical leap' (the second kind above), he tell the class specifically that he is doing so. For example "this step requires some proof, but it is a purely mathematical one which we will not delve into."

You can of course also ask him for a textbook with more mathematical rigor.

11

I hope the OP won't be too offended if I say that this question seems to show an immature understanding of the meaning of rigor and the relationship between different academic disciplines. As an illustration, consider the following problem, which could be used as an exam question on a freshman physics or calculus exam.

A uniform rod with mass per unit length b is initially upright and at rest in a gravitational field g. At t=0, the rod is released. At a later time t, find the rate at which mass flows past a horizontal surface passing through the rod.

Those of us who are physicists or mathematicians can easily find "the" answer, which is bgt.

Now suppose we want to make this a little tougher so we can use it as an interview question for a potential TA. We state the question, but now we ask specifically for a high level of rigor in the answer.

If the field is math, a good answer might be something along the following lines. The solution of the problem involves a derivative. One way of defining a derivative is as a limit, and limits are in turn defined using epsilons and deltas. Here's a rigorous epsilon-delta proof that the limit we're talking about does converge.

Now suppose the field is physics. (I'm a physicist.) An example of a nice, rigorous answer would be one in which the interviewee explained why the observable we're talking about cannot possibly converge to the expression bgt. A sufficient argument for nonconvergence would be to point out that the rod is made of atoms, so the motion of mass across a horizontal line starts to look discrete once we get down to a certain scale. (An even nicer answer might focus on effects that might be more practically observable. For example, when the support of the rod is released, the disturbance travels outward through the rod at the speed of sound, not instantaneously.)

Both of these are rigorous approaches to knowledge, but they are different notions of rigor. One emphasizes the internal self-consistency of mathematics. The other emphasizes the careful consideration of how mathematical models relate to reality, which is more complicated.

  • 3
    Perhaps you run in different mathematical circles than I do, but I don't think I've ever met a mathematician who would consider the definition of a derivative to be a useful part of a rigorous solution to the problem you describe. Rigor in mathematical physics isn't about epsilons and deltas for their own sake; it's about saying very clearly what you're trying to calculate, and arguing very convincingly that what you're doing is going to calculate it. – Vectornaut Sep 26 '15 at 19:02
  • If you haven't done so already, I would encourage you to poll some mathematical physicists around you about what they'd consider a "rigorous" answer to the problem you posed. I, for one, would be curious about the results. – Vectornaut Sep 26 '15 at 19:03
  • @Vectornaut: I think you're missing the point. The point is not that the rigorous underpinnings are nontrivial to practitioners of either field. They are trivial to both, and that's why I chose to dramatize the story by making it a job interview for a TA position -- so that we would have some motivation to ask someone to consider the trivial foundational issues. The point is that a practitioner of discipline A, using the correct notion of rigor for A, says that foo converges to bar. Meanwhile the person in field B says that foo does not converge to bar. They're both right. – Ben Crowell Sep 27 '15 at 1:39
  • 3
    But, in your story, the practitioner of math is not using the correct notion of rigor for math. That's what I'm trying to say: if you want your story to communicate the idea it's supposed to communicate, you should rewrite it so that the practitioner of each discipline is using the correct notion of rigor for that discipline. – Vectornaut Sep 27 '15 at 18:26
10

In theoretical physics, for various reasons, standards of mathematical rigor tend to be looser than they are in math. Individual physicists' preferences vary widely, however. In my experience, which seems to be echoed by Pete L. Clark's, many physicists tend to default to a looser standard of rigor while teaching, so your lecturers may or may not think about their course material at a much tighter level of rigor than the one they present it at.

You're definitely not alone in being frustrated by leaps of mathematical faith in physics lectures, and spending a great deal of time trying to fill them in. Here are some things I'd recommend doing to help deal with this, based on my own experience.

  • Do try talking to your teacher outside of class about mathematical gaps that confused you. You may find that your teacher knows exactly how to fill them in, and simply omitted the details from their presentation in class.

  • Do seek out other mathematically-minded people at your university, especially more experienced people, and talk to them about the things that confused you. As Pete L. Clark notes, many mathematically-minded physicists (and physicsy-spirited mathematicians!) have a private stash of rigorous insight into the less rigorous parts of a typical physics class, built up over years of experiences like yours. At some universities, the math department can be a gold mine of knowledge like this.

  • As a corollary, do write down your own work when you fill in the gaps yourself! Someday, the three weeks you spent proving that series converges might save someone else three weeks of trouble.

  • Do remember that not everything in physics has been formulated rigorously, and some topics are notoriously resistant to mathematical formalization. When you're confused by reasoning used in a physics class or the physics literature, it can be hard to tell whether you've encountered a small crack that can be paved over with a few hours of thought, an big gap that can be bridged using sophisticated techniques hidden in some corner of the math literature, or a gaping chasm that people have tried and failed to cross for decades. This is another reason talking to more experienced people can be helpful.

On the other hand, here are some things I'd recommend not doing.

  • Don't think of gaps in mathematical reasoning as mistakes, especially when you're talking to other people about them. This doesn't match the way most physicists approach mathematical reasoning, and it can turn your conversations unpleasantly confrontational.

  • If you've tried bringing your confusions to your teacher after class, and they've been consistently unable to help you, don't keep asking, especially if they seem annoyed by your problems. Your teacher may just prefer a looser standard of rigor than you, and there's nothing you can do about that. Seek out other sources of help instead.

  • Don't ask about leaps of reasoning during class. If your teacher doesn't know how to fill them in, nothing is gained. If your teacher does know how to fill them in, that means they've made a conscious decision not to, so they might prefer to talk to you outside of class.

  • Don't feel responsible for filling the mathematical gaps in your physics classes. In the comments here, people have said that "students can (and should) check that the claims made by their physics profs do indeed hold," and that "it's common to learn these concepts and their proofs rigorously in the math class." In my experience, those things just aren't true. You'll hit problems that you don't have the tools to resolve, and you'll hit problems that nobody has found the tools to resolve. Your confusion is not your fault.

  • Don't feel like your teachers are responsible for filling the gaps either. They're just doing physics as physics is generally done, and sometimes as it has to be done.

  • Don't spend too much time and energy trying to fill the gaps. Pancaking yourself against the far wall of the canyon a few times is okay, but at some point it's best just to walk away. You may come back later and discover that you've gained the tools and knowledge you need to get over, or that there's a bridge just a few miles away, or that getting over isn't likely to happen any time this century.

  • But, with that said, don't stop looking for more rigorous and less confusing ways to understand physics. Efforts to shore up the mathematical foundations of physics have proven very worthwhile in the past, and I firmly believe that they'll keep proving worthwhile in the future. They may feel thankless, but they're not worthless, and I think they're great things to read about and think about when you have the time and energy to spare.

I hope at least some of this advice is helpful for you. If you ever bring your mathematical physics troubles to Math.SE, I hope I'll see your question, and I hope I'll have the time and the knowledge to help answer it.

  • 2
    This is a great answer. – Pete L. Clark Sep 26 '15 at 23:04
-1

Generally, you better assess yourself well enough to call out a lecturer publicly, otherwise, there's plenty of opportunities to make corrections privately. That is generally the more politically-correct path. Beyond that, you should know that by choosing to address it publicly, you are (consciously or not) engaging in a battle of power. Such a battle can have positive or negative outcomes.

More specifically, I have questions for you: why would a lecturer have to prove that a series converges? Further, shouldn't multi-dimensional integrals always have the same answer regardless of the path? Otherwise, there's a deeper problem in the formulation of the expression (like including terms from a domain that doesn't belong there).

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.