Pure mathematics problem solving and relevance to theoretical physics

[rhotonsix]

TL;DR

How relevant is it to study and work through problems in a book like Rudin's analysis to build a mathematical foundation for studying theoretical physics

I am recreationally self studying physics and math with the ultimate goal of understanding theoretical/mathematical framework of classical mechanics, general relativity, and quantum mechanics (in addition to the actual physics). The question I have is how deep in the realm of pure math does one need for such a goal. For example I am working my way through baby Rudin, and reading and working out his proofs in the text is not an issue, but a number of the end chapter problems are extremely challenging. How useful is it if my goal is to study theoretical aspects of physics to spend a lot of time solving the problems in a book like Rudin’s? Would my time be better spent on texts like Hubbard’s vector calculus/differential forms book which strike a balance between the pure (proves all the major theorems of multivariable analysis in text rigorously) and applied (treats computation as well)? I am in no rush and have no problem spending the time to tackle the Rudin level problems but just unsure if it is counterproductive to my goal, which is to build a decent mathematical foundation for studying theoretical physics. Thanks in advance.

 

[PeroK]

Summary:: How relevant is it to study and work through problems in a book like Rudin's analysis to build a mathematical foundation for studying theoretical physics

I am recreationally self studying physics and math with the ultimate goal of understanding theoretical/mathematical framework of classical mechanics, general relativity, and quantum mechanics (in addition to the actual physics). The question I have is how deep in the realm of pure math does one need for such a goal. For example I am working my way through baby Rudin, and reading and working out his proofs in the text is not an issue, but a number of the end chapter problems are extremely challenging. How useful is it if my goal is to study theoretical aspects of physics to spend a lot of time solving the problems in a book like Rudin’s? Would my time be better spent on texts like Hubbard’s vector calculus/differential forms book which strike a balance between the pure (proves all the major theorems of multivariable analysis in text rigorously) and applied (treats computation as well)? I am in no rush and have no problem spending the time to tackle the Rudin level problems but just unsure if it is counterproductive to my goal, which is to build a decent mathematical foundation for studying theoretical physics. Thanks in advance.

In general, you are better off studying vectors, calculus, differential equations and using mathematics as a tool. It doesn't hurt to know rigorous mathematics, but it's a big investment in time.

How much physics do you already know?

 

[rhotonsix]

Thanks for the reply. Makes sense. I took a couple of courses as an undergrad in mechanics and EM but its been 15 years and my career is completely unrelated so I am going back and reviewing - Taylor and Morin for mechanics, Griffith for EM. Again this is purely recreational so I am in no rush, just planning on taking my time building up my foundations on both the physics and math side of things and enjoy it as much as I can.

 

[hutchphd]

If you understand well the Mathematics used in Griffiths EM you will have a very solid foundation. That encompasses the trinity of subjects mentioned by @PeroK.

For me it is easier to get enthused by the physics before I get confused by the maths.

.

 

[nucl34rgg]

For a general understanding of physics, Rudin probably isn't necessary. It certainly won't hurt you, but it won't really help you. But if you are specifically focusing on theoretical physics, I would think you would eventually want a strong understanding of real analysis, topology, differential geometry, representation theory, and algebra. Theoretical physicists basically end up with the equivalent understanding of perhaps at least a mathematics BS or even an MS. It' not uncommon these days for entering graduate students in theoretical physics to have had a math BS as well and to have taken some analysis courses.

I'd like to add something. I loved Griffiths' books, but I do not recommend them without a solid understanding of PDE. You should read Haberman's PDE book before attempting Griffiths in my opinion. Haberman really helps you understand the mysterious mathematical techniques presented in Griffiths with much more clarity. Griffiths kind of rushes through them.

 

[Frabjous]

Physics uses a lot of applied mathematics. So it is an open question on what the proper level of understanding of pure mathematics is. I think that most would agree that you do not need to go beyond an undergraduate level of understanding of pure math. Of course, you are doing this for fun, so follow your heart.

The flip side of the question is how much physics do you want to learn? The level of mathematical sophistication of a senior undergraduate understanding is less than what is needed for a graduate level understanding.

The problem is that physics uses a lot of different types of math. If you try to learn it all beforehand, you’ll never get to the physics. My advice is decide on the physics topic that you want to learn and then figure out what math you need to learn for it. Repeat.

BTW, there are books that try to split the difference . Off the top of my head, Flanders on Differential Forms, Sommerfeld on Partial Differential Equations, Burke on Applied Differential Geometry.

 

[rhotonsix]

Thank you all for the helpful answers!

 

[Frabjous]

Sometimes older math books are more applicable to physics. For example Modern Analysis by Whittaker and Watson.

 

[nucl34rgg]

Sometimes older math books are more applicable to physics. For example Modern Analysis by Whittaker and Watson.

Loomis and Sternberg's calculus book is a good example of an old school rigorous math book with a lot of physics insights as well.

 

[Frabjous]

Given your mathematical bent you might enjoy “Invitation to Classical Analysis” by Duren.

From the intro:

This is a book for undergraduates. To be more precise, it is designed for students who have learned the basic principles of analysis, as taught to undergraduates in advanced calculus courses, and are prepared to explore substantial topics in classical analysis. And there is much to explore: Fourier series, orthogonal polynomials, Stirling’s formula, the gamma function, Bernoulli numbers, elliptic integrals, Bessel functions, Tauberian theorems, etc. Yet the modern undergraduate curriculum typically does not encompass such topics, except perhaps by way of physical applications. In effect the student struggles to master abstract concepts and general theorems of analysis, then is left wondering what to do with them.

It was not always so. Around 1950 the typical advanced calculus course in American colleges contained a selection of concrete topics such as those just mentioned. However, the development could not be entirely rigorous because the underlying theory of calculus had been deferred to graduate courses. To remedy this unsatisfactory state of affairs, the theory of calculus was moved to the undergraduate level. Textbooks by Walter Rudin and Creighton Buck helped transform advanced calculus to a theoretical study of basic principles. Certainly much was gained in the process, but also much was lost. Various concrete topics, natural sequels to the abstract theory, were crowded out of the curriculum.

The purpose of this book is to recover the lost topics and introduce others, making them accessible at the undergraduate level by building on the theoretical foundation provided in modern advanced calculus courses.

 

[rhotonsix]

Thank you all for the replies and book recommendations. I do have Loomis/Sternberg and have enjoyed the chapters I have read in that text - I will definitely read further and check out the other references you all suggested.