Courses Math Course Advice -- Harvard freshman planning to double major in physics and mathematics

Click For Summary
The discussion centers on a Harvard freshman's dilemma between taking Math 25B, a proof-based real analysis course, or Math 55B, which covers topology and complex analysis at a faster pace. The student is currently completing Math 25A and aims to double major in physics and mathematics, with a focus on high-energy theory and mathematical physics. Key points include the recognition that Math 25B is less intensive and allows for deeper understanding, while Math 55B, despite its difficulty, covers more advanced material relevant to the student's research interests. The student contemplates the trade-off between mastering foundational concepts in 25B and the opportunity to accelerate their studies with 55B. Concerns about the fast pace of Math 55B and its superficial treatment of real analysis are discussed, alongside the importance of mathematical maturity and the potential impact on the student's GPA.
  • #31
mathwonk said:
That actually might be a good plan. Joe is really outstanding, certainly a superstar. And the only topology in 55b is topology that is oriented towards complex analysis, so he might cover some of it.
Joe told me that with regard to topology 113 will be entirely self-contained -- all the necessary topology will be covered in the course.
mathwonk said:
You might actually learn more this way, (in 25b + 113), since my instinct is that 55b is now designed as sort of a survey course, just sampling many subjects quickly.
Yes, that certainly seems to be the case according to the reviews of past students -- a lot of material covered very fast and quite superficially. Although Auroux would beg to differ...
mathwonk said:
The main thing I learned in 55a is expressed in the first paragraph of the introduction to chapter 8 of Dieudonne'. Namely, the derivative of a function at a point, is a linear map approximating ("tangent to") that function near the point. Hence the derivative of a composition of functions is the composition of their linear approximations.
Could one see this as an intuitive explanation for the chain rule?
mathwonk said:
By the way, one of my students was a phi beta kappa physics major at Harvard some decades ago, and took the math 25 sequence (from the famous superstar topologist Raoul Bott), so math 25 may serve you also quite well.
Out of curiosity, do you know what math 25 was like back in your days?
mathwonk said:
You might well discuss your choices with Professor Harris, who is very helpful and friendly.
I have already talked to him! He said that all the options are very good. In his opinion, 55b has a slight edge over the 25b + 113 combo in that it covers algebraic topology in addition to real and complex analysis. At any rate, he said that I should shop all the courses and decide which ones I feel most comfortable in, which is probably what I'll do.
 
Physics news on Phys.org
  • #32
Sounds as if you have covered all the bases you can, and are well prepared to choose.

To your questions: yes that remark of Dieudonne' is exactly an explanation of why the chain rule is a completely natural result, properly formulated as a statement about maps. The Inverse function theorem also has a similar statement: If f is a smooth function from R^n to R^n, taking say 0 to 0, and if the derivative at 0, is invertible as a linear map, then f itself is invertible as a smooth map, on some neighborhood of 0. The implicit function theorem can also be stated: if the derivative f'(0) becomes a projection after a linear change of coordinates, then f also becomes a projection on some neighborhood of 0, after a smooth change of coordinates. (I strongly advise however also learning the usual old fashioned statement and applications of this theorem, for use in practice. In my case at least, after learning this abstract statement, I had no idea what Mumford meant when he said the implicit function theorem means, given an equation, you can use it to solve for some of the variables in terms of the other variables. )

In my day, there was no math 25. There was the 2 year 11/55 sequence, and the 3 year 1/20/105 sequence, for calculus, plus honors versions of 1/20/105. Math 11 "used" (i.e. recommended and totally ignored) Courant, later Spivak Calculus, Math 20 used a book like Angus Taylor, and 105 (or honors 105) used maybe Courant, or David Widder, while 55 was based on Dieudonne', or Apostol's Mathematical Analysis, later Loomis/Sternberg.
Apparently 25 was created as a middle ground between the two sequences, more sophisticated than 20/105, but less drastic than 55. There was also the changing situation in which virtually all strong high school students began to arrive with some calculus, whereas earlier very few had it available in high school.

AP calculus in high school caused a problem for teaching college math, since it creates a misconception in the students that they should be entitled to "advanced placement". But many high school "AP" students are not even prepared to skip one non honors calc course in college. If they try, one has these poorly prepared students in second semester calc, but not knowing the first semester well. This causes many good students to actually fail out of second semester or second year calculus. One way to adapt is by lowering the level of the college courses to match the weaker level of preparation.

"Advanced placement" also orients these students poorly, since they go from being honors high school students to taking advanced but non honors college classes. Thus their peer group is wrong.

At some schools, like UGa and Chicago, the old math 11 style "Spivak" course was kept for entering honors students, followed by an advanced version, based (at UGa) on Ted Shifrin's book:
https://www.amazon.com/Multivariabl...lus-Manifolds/dp/047152638X?tag=pfamazon01-20

Other schools like Stanford and Harvard, dropped the beginning Spivak course and plunged bright honors students right into a 55 style course, often with disastrous results. Rethinking these offerings perhaps resulted in the creation of math 25.
The physics major mentioned earlier, who started Harvard in math 25, had already prepared in high school with AP calculus followed by a proof based course from Marsden and Tromba's Vector Calculus book, augmented by an introduction to differential forms, (and taught by a college professor, i.e. me). He remarked that he "could not have survived" Bott's math 25 without the preparation in differential forms. (Check out the beautiful (advanced) book by Bott-Tu: Differential Forms in Algebraic Topology).

(If you choose or evaluate your courses based on the level of the professor, you begin to realize what it means to take beginning calculus as an AP course in high school, sometimes from someone who was not even a math major, but had only taken calculus in college.)

We have talked only about sources for advanced calculus, but I also recommend getting hold of a copy of Michael Spivak's Calculus, as a resource for one variable calculus, done right; or maybe better Courant, or Apostol's Calculus, since those books mention physics, unlike Spivak.
Here are used copies of both volumes of Courant, for under $20 total, a bargain.
https://www.abebooks.com/servlet/SearchResults?an=Courant&cm_sp=SearchF-_-home-_-Results&ref_=search_f_hp&sts=t&tn=differential and integral

Oh yes, and this discussion reminds of the third aspect of a college class: 1) professor, 2) syllabus, and 3) classmates! In terms of stimulating classmates, math 55b should probably be favored.
But looking again at the syllabi, that 25b syllabus is already quite extensive, i.e. rigorous one and several variable calculus, and the profs are both stars. So, and obviously the choice is up to you not me, but 25b + 113 looks to me like a more reasonable learning experience, at least on paper. So I guess the only way to decide is to try them yourself and get a feel, as you said. Good luck!

By the way, the stuff on axioms for real numbers in the 25b syllabus is covered in Dieudonne' chapters 1, 2. (And it seems pretty ambitious to cover all of the foundations of real numbers in just one 25b lecture, unless it is review.) If you can read that and maybe also chapter 3, you can likely handle anything theoretical you will see in these courses. I.e. Dieudonne' is the hardest source to read of all I mentioned. Here is a used copy for around $32. This is an amazing book for content, but very demanding reading (no pictures!).
https://www.abebooks.com/servlet/Se...Results&ref_=search_f_hp&sts=t&tn=foundations
 
Last edited:
  • #33
mathwonk said:
Sounds as if you have covered all the bases you can, and are well prepared to choose.
In my day, there was no math 25. There was the 2 year 11/55 sequence, and the 3 year 1/20/105 sequence, for calculus, plus honors versions of 1/20/105.
What I find interesting is that back in the day one would take 55 after taking a more elementary course like 11. Nowadays, 55 is only open to freshmen and so you are supposed to take it right out of high school which for many, including myself, is not an easy thing to do.
mathwonk said:
He remarked that he "could not have survived" Bott's math 25 without the preparation in differential forms. (Check out the beautiful (advanced) book by Bott-Tu: Differential Forms in Algebraic Topology).
Oh wow, 25 must have been watered down then, because nowadays differential forms are only taught in 55b. And even then, most of the class (from what I have been told) does not really understand them, because at the level of 55b, one lacks the mathematical tools to fully grasp the concept -- this kind of machinery is taught in a class like 132 (differential topology).
mathwonk said:
We have talked only about sources for advanced calculus, but I also recommend getting hold of a copy of Michael Spivak's Calculus, as a resource for one variable calculus, done right; or maybe better Courant, or Apostol's Calculus, since those books mention physics, unlike Spivak.
Here are used copies of both volumes of Courant, for under $20 total, a bargain.
https://www.abebooks.com/servlet/SearchResults?an=Courant&cm_sp=SearchF-_-home-_-Results&ref_=search_f_hp&sts=t&tn=differential and integral
Speaking of calculus/analysis books, where does Rudin's Principles of Mathematical Analysis fit into your classification? Is it of a similar calibre to Spivak and Apostol? You don't seem to mention it. Is there a reason why you prefer those other books to Rudin? I'm asking because 55b lists Rudin as the main (and only) source on real analysis.
mathwonk said:
Oh yes, and this discussion reminds of the third aspect of a college class: 1) professor, 2) syllabus, and 3) classmates! In terms of stimulating classmates, math 55b should probably be favored.
Yes, I can confirm that 55 has a better math community than 25. My first week in 55a was sufficient for me to realize that. On the other hand, in 55 one is surrounded by mathematical geniuses -- for example, this year there is one guy taking 55, 114 (integration and measure theory) and a graduate course on complex analysis -- and as a result it's quite easy to feel intimidated.
mathwonk said:
But looking again at the syllabi, that 25b syllabus is already quite extensive, i.e. rigorous one and several variable calculus, and the profs are both stars.
Yes, the 25b syllabus is definitely very solid.
mathwonk said:
So, and obviously the choice is up to you not me, but 25b + 113 looks to me like a more reasonable learning experience, at least on paper. So I guess the only way to decide is to try them yourself and get a feel, as you said. Good luck!
I think I would probably learn more from 25b + 113. That said, there is also the hypothetical option of doubling up on physics and taking two physics courses instead of one, in which case I would most likely have to limit myself to a single math course. I guess I need to evaluate my priorities and decide which of the two subjects -- math or physics -- I want to zoom in on this spring.
mathwonk said:
By the way, the stuff on axioms for real numbers in the 25b syllabus is covered in Dieudonne' chapters 1, 2. (And it seems pretty ambitious to cover all of the foundations of real numbers in just one 25b lecture, unless it is review.) If you can read that and maybe also chapter 3, you can likely handle anything theoretical you will see in these courses. I.e. Dieudonne' is the hardest source to read of all I mentioned. Here is a used copy for around $32. This is an amazing book for content, but very demanding reading (no pictures!).
https://www.abebooks.com/servlet/Se...Results&ref_=search_f_hp&sts=t&tn=foundations
Thank you so, so much @mathwonk for all this advice and all the links! I'll definitely check them out over winter break.
 
  • #34
In 1960, math 11a covered roughly what is now on the syllabus for the first half of math 25b. It included axioms for the reals (handed out on day one), convergence of sequences and series, (real and complex) power series (applied to define exponential and trig functions), existence of maxima and minima of continuous real valued functions on closed bounded intervals, intermediate and mean value theorems, FTC, and probably more. I don't remember much from math 11b , but I do seem to recall coverage of vector spaces, including Hilbert and pre-Hilbert spaces, possibly Fourier series, and applications to differential equations.

Math 55 today is indeed very challenging for freshmen, but there some freshmen who have extremely strong preparation. Some years back when math55 was offered by Wilfried Schmid, the note taker had taken all the honors level college calculus courses at UGa, including the Spivak Calculus style course and the Shifrin multivariable calculus course, plus more advanced courses, while still in high school, and then took math 55 at Harvard, very well prepared. As I said, my physics student had taken one and several variable calculus in high school, including vector calculus from Marsden and Tromba, and had learned to compute with differential forms, and then even he took 25 rather than 55. It is not entirely clear to me what preparation is expected, or usual, for success in 55.

As to learning differential forms, there was a thread here on PF devoted to that topic some years back, and they went through a very nice book by David Bachman, available from him free online:
https://faculty.washington.edu/seattle/physics544/2011-lectures/bachman.pdf

As for Rudin's Principles of analysis book, it is famous for being very precise but very unmotivated, so I never recommend it for learning. But analysis professors love to recommend it, so you should take a look, maybe it will work for you. I prefer books by Spivak, Apostol, Berberian, Fleming, Lang, Simmons, and although quite difficult I admire Dieudonne'.


Here are some rather nice, and very down to earth, notes on differential forms, and their use in calculus, by Donu Arapura, that make them seem quite reasonable:
https://www.math.purdue.edu/~arapura/preprints/diffforms.pdf

and here is a very nice and accessible book on the topic, aimed at physical scientists. (I got over my own fear of differential forms in a little article by Flanders where he just taught me to calculate with them. being able to manipulate them made them seem less scary. Actually all they are is [combinations of and families of] determinants.)
https://store.doverpublications.com...Yftbzdo08qK2Zz3RT_1EAaXU51xIG5L5q1h1iHvx9Ordi

i.e. they are (linear combinations of) families of operators which, at each point, look at an ordered finite set of vectors and spit out the oriented volume of the parallelepiped those vectors (or certain of their components) span.

But don't be intimidated by wondering what they "are", just learn how they behave, i.e. learn to add, multiply, differentiate, and pull them back. When they give us axioms for real numbers these also only tell us how numbers behave, not what they are.

On the topic of the student body in math 55: yes one will naturally feel intimidated at times, maybe most of the time. I definitely felt that way when there. But the good part is that, if we can ignore that and not worry about how we compare to these geniuses, it turns out that we ourselves get stronger by being around them, as long as we do actually interact with them and open up to what they can offer us. I.e. when I later went elsewhere, it seemed my own level had improved from being around these exceptional people. In yoga, they speak of the aspect of someone who, although not a saint, has "been with saints". And they may be saying the same about you: "Here is this student who is taking these high powered math major courses and simultaneously majoring also in physics!"

If you want to test how much linear algebra you have learned in 25a, here is a revised set of the notes on introductory linear algebra I originally wrote for fun one Christmas break, first in 15 pages to see how short I could make it, and then expanded to 125 pages, as here. Determinants are only summarized. (The 15 page version is also on my website, linked above in post #27, if you are curious.)
https://www.math.uga.edu/sites/default/files/laprimexp.pdf

and here is another shorter (69pages) set of linear algebra notes from a course, with a complete treatment of determinants at the end.
https://www.math.uga.edu/sites/default/files/inline-files/4050sum08.pdf
 
Last edited:
  • #35
Quantum55151 said:
He says that I could definitely give 55B a try, but in his view, 55B is too superficial in its treatment of the material, especially of real analysis.
But isn't that okay with you, since you intend to be a physicist and not a mathematician?
 
  • #36
@mathwonk Sorry for the delay in my response. I was preparing for finals.
mathwonk said:
It is not entirely clear to me what preparation is expected, or usual, for success in 55.
Auroux writes in the course syllabus that no formal knowledge of linear algebra, group theory or analysis is required, but a familiarity with proof-writing and abstract reasoning as well as a commitment to a fast-paced course is. What this actually means in practice is a separate question.
mathwonk said:
As to learning differential forms, there was a thread here on PF devoted to that topic some years back, and they went through a very nice book by David Bachman, available from him free online:
https://faculty.washington.edu/seattle/physics544/2011-lectures/bachman.pdf
I'll check out, thanks!
mathwonk said:
As for Rudin's Principles of analysis book, it is famous for being very precise but very unmotivated, so I never recommend it for learning. But analysis professors love to recommend it, so you should take a look, maybe it will work for you. I prefer books by Spivak, Apostol, Berberian, Fleming, Lang, Simmons, and although quite difficult I admire Dieudonne'.
Do you know anything by Elements of Classical Analysis by Marsden and Hoffman? This is the book used in 25b.
mathwonk said:
On the topic of the student body in math 55: yes one will naturally feel intimidated at times, maybe most of the time. I definitely felt that way when there.
Even though you were the oldest student?
 
  • #37
Muu9 said:
But isn't that okay with you, since you intend to be a physicist and not a mathematician?
True. At the same time, I still want to major in math and do physics research that is more on the mathy side. So I do want to get a solid grasp on the material.
 
  • #38
I don't know the book by Marsden/Hoffman. Reviewers on amazon say its helps that they teach n dimensions starting from 2,3 dimensions., which sounds user-friendly.


oldest ≠ best. when I was a freshman in math 11, there was another freshman who lived across the hall from me who was a year or 2 younger, and taking math 55. he occasionally helped me in math 11. I guess I wasn't so much intimidated by these very gifted students as impressed. The intimidation factor was greater later as a postdoc interacting with the professors.
 
Last edited:
  • Like
Likes berkeman and Quantum55151
  • #39
please let us know what you choose and how it goes. good luck!
 
  • #40
Quantum55151 said:
Hi everyone!

I would like to get advice concerning my math course for the spring semester.

For context, I am a Harvard freshman planning to double major in physics and mathematics with the long-term goal of doing research in high-energy theory and/or mathematical physics. I am currently finishing math 25A, a proof-based linear algebra course. For the spring semester, I have the option of either taking 25B, a proof-based real analysis course, or moving up to Harvard's infamous math 55. In the spring, math 55B covers topology (both point-set and algebraic) and complex analysis with a brief two weeks of real analysis in between.

Now, 25B is an objectively easier class, because it attempts to cover much less material in one semester than does 55B and also assumes less mathematical maturity on the part of students. For me at least, this might translate into a better understanding of real analysis than what I would gain from 55B where the professor tries to speed run through Rudin over the course of...2 weeks lol. At the same time, based on what I've read (and please correct me if I am wrong), I don't think real analysis is particularly useful for theoretical physics (at least not the kind that is taught in 25B; integration and measure theory might be a different story, but that is taught in a different class altogether). Topology and complex analysis, on the other hand, seem to be much more relevant to the kind of physics that I want to do. The other nice thing about 55B is that it would allow me to save time by basically knocking out three undergrad math courses in one semester which in turn would allow me to take more advanced undergrad classes and/or grad classes sooner. On the flip side, the trade-off will consist in how well I will actually learn the material in 55B as well as the level of difficulty of the class which, although a far cry from the stuff you'll read on 55's Wikipedia page, is nevertheless non-negligible.

What do you think? At any rate, I can always try out 55B for the first few weeks and then drop down to 25B if necessary. But I would still appreciate any advice so that I have a better idea of my spring plans and can plan out my winter studies accordingly.

P.S. I am attaching the syllabi for the two courses in case anyone wants to take a look.
Know yourself. You have way more math ability than I did. I took Physics 12A, bombed it, was grateful for a C, took Physics 1 w Paul Bamberg & aced it. (This was 1973)
Bamberg advised me about going on to Physics 112 and was insightful ( I didn't.)
I recommend you speak w him. He's a great teacher and straddles your departments.
 
  • #41
Quantum55151 said:
@mathwonk Sorry for the delay in my response. I was preparing for finals.

Auroux writes in the course syllabus that no formal knowledge of linear algebra, group theory or analysis is required, but a familiarity with proof-writing and abstract reasoning as well as a commitment to a fast-paced course is. What this actually means in practice is a separate question.

I'll check out, thanks!

Do you know anything by Elements of Classical Analysis by Marsden and Hoffman? This is the book used in 25b.

Even though you were the oldest student?

Quantum55151 said:
@mathwonk Sorry for the delay in my response. I was preparing for finals.

Auroux writes in the course syllabus that no formal knowledge of linear algebra, group theory or analysis is required, but a familiarity with proof-writing and abstract reasoning as well as a commitment to a fast-paced course is. What this actually means in practice is a separate question.

I'll check out, thanks!

Do you know anything by Elements of Classical Analysis by Marsden and Hoffman? This is the book used in 25b.

Even though you were the oldest student?
I had Hoffman as an instructor many years ago. Extremely knowledgeable person. He would also give us expanded notes, and what he would believe are shortcommings in the text, and how, he planned to update it in a later edition.
I do not have the notes, but I will ask around.

I would rank it as a slightly easier version of Apostol. It isnt bad. It has a few difference. Ie., proofs are left to the end of section [or maybe chapter]. Good because it allows you to work on the proof without peaking. Annoying, for review. Since you have to flip through pages.

I went to a state school [low ranked] and found it insightful. This may or may not be your experience. You can always get a copy of Pugh if you find Marsden/Hoffman too easy.

I believe it covers about the same topics as Rudin [ i only glanced at early in my education].

I prefered Marsden/Hoffman complex analysis book much more.
 
  • #42
@Quantum55151: I am curious as to your eventual class choices and your assessment of the experience now.
 

Similar threads

  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 32 ·
2
Replies
32
Views
2K
  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 40 ·
2
Replies
40
Views
6K
  • · Replies 9 ·
Replies
9
Views
2K
  • · Replies 6 ·
Replies
6
Views
3K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 7 ·
Replies
7
Views
3K
  • · Replies 7 ·
Replies
7
Views
2K