Should I Become a Mathematician?

In summary, to become a mathematician, you should read books by the greatest mathematicians, try to solve as many problems as possible, and understand how proofs are made and what ideas are used over and over.
  • #36
Becoming a mathematician part4) Starting College

Becoming a mathematician, part 4) College training.

I suspect it does not matter greatly which college you go to, as they all have their strengths and weaknesses. Places like Harvard or Stanford or Berkeley offer famous lecturers on a high level, incredibly advanced courses, and brilliant highly competitive students. For many of us, this can be more intimidating than inspiring. And often the famous professors are simply unavailable for conversation outside of class. In the early 60’s at Harvard, I found the lectures were wonderful, if I got the best professors, and then they walked out and I never saw them again until next time. Office hours were minimal and if I tried to see some of them, they were frequently busy or uninterested. Even intelligent questions in class seemed as likely to be met with sarcasm as a helpful answer. I suspect things have changed now with people like Joe Harris and Curt McMullen there, who are great teachers as well as researchers, and who enjoy students. Of course there were outstanding teachers like Tate and Bott there in the old days too, but not everyone was like them. As a result, I had to go away and get back my enthusiasm for math at a more supportive place.

It is helpful to go somewhere where you will enjoy your time, enjoy the courses and the other students, and get help from professors who think students matter. Today this is more common everywhere, even at famous universities, than it was long ago, but ask around among the student body. And be prepared to work very hard. Some if not most of my own undergraduate frustrations could have been lessened, possibly solved, by better study habits.

As to what courses to take, this is tricky and complicated by the almost worthless AP preparation most kids get today in high school. In general an AP class is a class taught by someone with nowhere near the training or understanding of a college professor, although they may be a fine teacher. But to expect a calculus course taught by an average high schol math teacher to substitute for a honors introduction to calculus taught by Curt McMullen or Wilfried Schmid or Paul Sally, is ridiculous. Nonetheless, so many students have bought this ridiculous idea that Harvard and Stanford do not even offer an honors introduction to calculus anymore for future math majors. There simply are none out there who have not had AP calculus in high school. Thus the student entering from high school is faced with beginning in one of many choices of several variable calculus courses. The most advanced one, the one taught a la Loomis and Sternberg, realistically requires preparation in a very strong one variable course a la Apostol, but which Harvard does not itself offer. So the only students prepared to take it are those elite ones coming in from Andover or Exeter or the Bronx high school of science, but not the rest of us coming in with our inadequate AP courses from normal high schools.

Thus the jump from high school to college has been made harder by the existence of AP courses. So in my opinion, even with AP calculus preparation, it is often helpful for a prospective mathematician to try to begin college in an introductory, but very challenging, one variable calculus course, modeled on the books of Spivak or Apostol, if you can find them. These do exist a few places, such as University of Georgia, and University of Chicago, which still offer beginning Spivak style calculus honors courses. To quote the placement notice from Chicago: “The strong recommendation from the department is that students who have AP credit for one or two quarters of calculus enroll in honors calculus (math 16100) when they enter as first year students. This builds on the strong computational background provided in AP courses and best prepares entering students for further study in mathematics.”

(I am not positive, but I assume that 16100 is the spivak course. But do your own homeworkl to be sure.)

The point is that AP preparation provides no theoretical understanding, so plunging students into advanced and theoretical calculus courses of several variables, as they do at Harvard and Stanford, by beginning in Apostol vol 2, or Loomis and Sternberg, without background from Apostol volume 1 or Spivak, is academic suicide even for most very bright and motivated students.

If you go to a school where there is no Spivak or Apostol vol. 1 type course, where the calculus preparation is from Stewart, or some such book, you are perhaps getting another AP course, only in college. Then you have to choose more carefully. Many such college courses will indeed be no more challenging than a high school AP course, and should not be repeated. Just ask the professor. They know the difference, and will help you choose the right level course. Either get in an honors section, or an advanced course suitable to your background. And join the math club. Try to find out who the best professors are, and do not be scared off if weak students say a certain professor is tough. You may not think so if you are a strong student. Once you get there, try to sit in on courses before taking them, to see which professors suit you. Student evaluations are notoriously hard to interpret correctly. The professor with the worst reputation among students, Maurice Auslander, was in my grad school days at Brandeis my absolute favorite professor. He cared the most, offered the most, and taught us the most. He also worked us the hardest.

Once you get a semester or two under your belt, it will get easier to find the right class, as hopefully the colleges own courses prepare you for their continuations, although this is not guaranteed! There is no way to force one professor to included everything the enxt one expects, nor to exclude material he/she loves that is outside the curriculum. Do your own investigating. Ask the professor what is needed for his/her course and try to get it on your own if necessary. After leaving the honors program temporarily as an undergraduate, I got back in by studying on my own over the summer from an advanced calculus book (David Widder), to make up my theoretical deficiencies and survive the next course.

Everyone should study calculus, linear algebra, abstract algebra, ode, and some basic topology. If you have no background in proofs from high school, you will need to remedy that as soon as possible. It is best to do this before entering, even if they offer a “proofs and logic” course. Such courses are often offered to junior math majors, whereas they are needed to understand even beginning courses well. For this reason it is extremely helpful to read good math books on your own that contain proofs. Today especially it is important to know some physics even if if you only plan to do math. Much of the inutition and application of math comes from physics. Even if you only want to do number theory, sometimes viewed as the purest and most esoteric branch of math, many of the deepest ideas in number theory come from geometry and analysis and even statistics, so nothing should be skipped. Work hard, read good books, seek good teachers, and try to have fun. College is potentially the most exciting and fun time of your life, and the one where, believe it or not, you have the most freedom and free time.
 
Physics news on Phys.org
  • #37
Wow. You would do a superb PhD if you have the inclination, but as you are already earning a living that would be a sacrifice.

You have the innate power and creativity of a PhD level mathematician. This is unusual with only a BS.
When I thought I was going to look for a programming job, my plan was to go back to school and learn more math.

But since I'm actually employed as a mathematician (and have become fairly good at self-study), I don't feel as much need. OTOH, my employer will pay for some full-time schooling (both the classes, and giving me my full pay!), so I really ought to take advantage of it. My buddies keep trying to tell me to go and get a masters in logic. :smile:
 
  • #38
Becoming a mathematician part5) some good books

Some recommended undergraduate books for future mathematicians.

Introductory calculus.
1. Calculus (ISBN: 0521867444)
Spivak, Michael Bookseller: Blackwell Online
(Oxford, OX, United Kingdom) Price: US$ 53.66
Shipping within United Kingdom:FREE

Book Description: Cambridge University Press, 2006. Hardback. Book Condition: Brand New. 3Rev ed. *** CONDITION NEW COPY *** TITLE SHIPPED FROM UK *** Pages: 672, Spivak's celebrated textbook is widely held as one of the finest introductions to mathematical analysis. His aim is to present calculus as the first real encounter with mathematics: it is the place to learn how logical reasoning combined with fundamental concepts can be developed into a rigorous mathematical theory rather than a bunch of tools and techniques learned by rote. Since analysis is a subject students traditionally find difficult to grasp, Spivak provides leisurely explanations, a profusion of examples, a wide range of exercises and plenty of illustrations in an easy-going approach that enlightens difficult concepts and rewards effort. Calculus will continue to be regarded as a modern classic, ideal for honours students and mathematics majors, who seek an alternative to doorstop textbooks on calculus, and the more formidable introductions to real analysis. Preface; Part I. Prologue: 1. Basic properties of mumbers; 2. Numbers of various sorts; Part II. Foundations: 3. Functions; 4. Graphs; 5. Limits; 6. Continuous functions; 7. Three hard theorems; 8. Least upper bounds; Part III. Derivatives and Integrals: 9. Derivatives; 10. Differentiation; 11. Significance of the derivative; 12. Inverse functions; 13. Integrals; 14. The fundamental theorem of calculus; 15. The trigonometric functions; 16. Pi is irrational; 17. Planetary motion; 18. The logarithm and exponential functions; 19. Integration in elementary terms; Part IV. Infinite Sequences and Infinite Series: 20. Approximation by polynomial functions; 21. e is transcendental; 22. Infinite sequences; 23. Infinite series; 24. Uniform convergence and power series; 25. Complex numbers; 26. Complex functions; 27. Complex power series; Part V. Epilogue: 28. Fields; 29. Construction of the real numbers; 30. Uniqueness of the real numbers; Suggested reading; Answers (to selected problems); Glossary of symbols; Index. Bookseller Inventory # 0521867444

2a. Calculus. Volume I. One-Variable Calculus, with an Introduction to Linear Algebra. Second Edition
Apostol, Tom M Bookseller: Paper Moon Books
(Portland, OR, U.S.A.) Price: US$ 20.00
Shipping within U.S.A.:US$ 4.50
Book Description: New York John Wiley & Sons, Inc. 1967., 1967. Fine. 666pp. Clean and bright book. No previous owner's markings. 2nd.Edition. Binding is Hardback. Bookseller Inventory # 068435

2b. Calculus. Volume II. Multi-Variable Calculus and Linear Algebra with Applications to Differential Equations and Probabil
Apostol, Tom M Bookseller: Paper Moon Books
(Portland, OR, U.S.A.) Price: US$ 20.00
Shipping within U.S.A.:US$ 4.50
Book Description: New York John Wiley & Sons, Inc. 1969., 1969. Fine. 673pp. Clean and bright book. No previous owner's markings. 2nd.Edition. Binding is Hardback. Bookseller Inventory # 068436

3a. Introduction to Calculus and Analysis (Volume I)
Courant, Richard; Fritz John
Bookseller: Harvest Book Company
(Fort Washington, PA, U.S.A.) Price: US$ 9.95
Shipping within U.S.A.:US$ 3.95
Book Description: Interscience Publishers/ New York 1965, 1965. First American Edition, 1st Printing Hardback in Decorated Boards. 661p. Very good condition. Very good dust jacket with one small closed tear and sunned jacket spine. Satisfaction Guaranteed. Bookseller Inventory # 515288

3a, alt. Introduction to Calculus and Analysis Volume 1 (ISBN: 0470178604)
Richard Courant
Bookseller: Frugal Media Corporation
(Austin, TX, U.S.A.) Price: US$ 10.00
Shipping within U.S.A.:US$ 3.70
Book Description: Wiley, John Sons. Hardcover. Book Condition: VERY GOOD. USED Ships within 12 hours. Bookseller Inventory # 873302

3b. Differential and Integral Calculus Volume 2
R. Courant
Bookseller: Pioneer Book
(Provo, UT, U.S.A.) Price: US$ 13.50
Shipping within U.S.A.:US$ 3.50
Book Description: Interscience Publishers, 1947. rebound Hard Cover Good. Bookseller Inventory # 481571

4. ANALYSIS 1
Lang, Serge
Bookseller: The Book Cellar, LLC
(Nashua, NH, U.S.A.) Price: US$ 39.99
Shipping within U.S.A.:US$ 4.00
Book Description: Addison-Wesley 1968., 1968. Fine in Good dust jacket; Light shelf wear to book. Heavy wear to DJ. 460 pages. Binding is Hardcover. Bookseller Inventory # 374309

5. Calculus of One Variable
Joseph W. Kitchen, Jr. Bookseller: Antiquarian Books of Boston
(Winthrop, MA, U.S.A.) Price: US$ 150.00 [sorry]
Shipping within U.S.A.:US$ 3.50
Book Description: Addison-Wesley Publishing, Reading, Mass., 1968. Hard Cover. Book Condition: Very Good. No Jacket. 8vo. xiii, 785 pages. Tightly bound and clean. No writing in book. The book also deals with plane analytic geometry and infinite series. Bookseller Inventory # 7620

also Honours Calculus*(ISBN: 0965521117) $24. from the author.
Helson, Henry
http://members.aol.com/hhelson/ [Broken]


Calculus of several variables.
6. CALCULUS ON MANIFOLDS: A MODERN APPROACH TO CLASSICAL THEOREMS OF ADVANCED CALCULUS
Spivak, Michael Bookseller: BRIDGEWAY ACADEMIC BOOKSTORE, ABA
(TAOS, NM, U.S.A.) Price: US$ 25.00
Shipping within U.S.A.:US$ 6.50
Book Description: W. A. Benjamin, NY, 1965. PAPERBACK COPY. Book Condition: Very Good. VERY GOOD CONDITION, PAPERBACK, 146pp. Bookseller Inventory # 001874

7. Mathematical Analysis: A Modern Approach to Advanced Calculus
Apostol, T. M. Bookseller: Textsellers.com
(Hampton, NH, U.S.A.) Price: US$ 12.50
Shipping within U.S.A.:US$ 3.50
Book Description: Addison Wesley, 1957. Book Condition: Good. Dust Jacket Condition: Fair. 8vo - over 7¾" - 9¾" tall. Hardcover, 559 pp. Notes, jacket has edge chips. Bookseller Inventory # 011916

8. Functions of Several Variables.
Fleming, Wendell H. Bookseller: Significant Books
(Cincinnati, OH, U.S.A.) Price: US$ 12.00
Shipping within U.S.A.:US$ 3.50
Book Description: 337 pp. Addison Wesley (1965) (Hardback) Good condition, ExLib. Glue Spot on cover. Bookseller Inventory # MATH10273

9. Advanced Calculus
Loomis and Sternberg
free download from Sternberg’s website.

Linear Algebra:
10. Linear Algebra : A Geometric Approach (ISBN: 071674337X)
Malcolm Adams, Ted Shifrin Bookseller: www.EMbookstore.com[/URL]
(Flushing, NY, U.S.A.) Price: US$ 67.98
Shipping within U.S.A.:US$ 3.25
Book Description: W. H. Freeman; (August 24, 2001), 2001. Book Condition: New. Free Delivery Confirmation! Brand New Hardcover, US Edition, Quality Paper Printed in USA. Bookseller Inventory # 071674337X-2

11. Linear Algebra.
Hoffman, Kenneth, & Ray Kunze Bookseller: Zubal Books
(Cleveland, OH, U.S.A.) Price: US$ 11.46
Shipping within U.S.A.:US$ 6.50
Book Description: Englewood Cliffs: Prentice-Hall 1965, 1965. 1st edition, fourth printing (1965) 332 pp., hardback, wear to spine & covers, previous owner's name to front free endpaper else textually clean & tight. Bookseller Inventory # ZB471098

Ordinary Differential Equations
12. Ordinary Differential Equations (ISBN: 0262510189)
V. I. Arnold Bookseller: A1Books
(Netcong, NJ, U.S.A.) Price: US$ 28.77
Shipping within U.S.A.:US$ 4.95
Book Description: Brand new item. Over 3.5 million customers served. Order now. Selling online since 1995. Few left in stock - order soon. Code: M20060602184422T0262510189. SKU: 0262510189-11-MIT. Bookseller Inventory # 0262510189-11-MIT

13. Lectures on Ordinary Differential Equations.
Hurewicz, Witold. Bookseller: Significant Books
(Cincinnati, OH, U.S.A.) Price: US$ 7.00
Shipping within U.S.A.:US$ 3.50
Book Description: Book Condition: Good condition, no dj. 122 pp. Wiley (1958 ) Hardback. Bookseller Inventory # MATH12978


Topology
14. First Concepts of Topology
Chinn, W. G. & Steenrod, N.e. Bookseller: aridium internet bookstore
(Cranbrook, BC, Canada) Price: US$ 8.32
Shipping within Canada:US$ 8.95
Book Description: SInger, 1966. Trade Paperback. Book Condition: Very Good. First Printing. Usual library markings in and out. non-circulating. very light use, clean crisp pages. edge rub/wear. A solid copy. Ex-Library. Bookseller Inventory # 010917

15. Differential Topology: First Steps
Wallace, Andrew Bookseller: Books on the Web
(Winnipeg, MB, Canada) Price: US$ 30.25
Shipping within Canada:US$ 5.50
Book Description: NY: W.A. Benjamin, 1968, 1968. paper bound, 1st edition, illustrated in colour, 130pp including bibliography and index. As new. Bookseller Inventory # 16779

16. An Introduction to Algebraic Topology
Wallace, Andrew H. Bookseller: BOOKS - D & B Russell
(Shreveport, LA, U.S.A.) Price: US$ 12.00
Shipping within U.S.A.:US$ 4.00
Book Description: Pergamon Press, New York, 1963. Book Condition: Very Good hard cover/ no dust. Octavo, 198 pp., Last name of prior owner inside front cover. One of a series of the International Series of Momographs in Pure and Applied Mathematics. Bookseller Inventory # 013208


Abstract Algebra.

17. Algebra (ISBN: 0130047635)
Artin, Michael Bookseller: DotCom Liquidators / DC 1
(Fort Worth, TX, U.S.A.) Price: US$ 44.50
Shipping within U.S.A.: US$ 3.50
Book Description: Bookseller Inventory # NA/DC8/T999/*114552

Abstract Analysis

18. Foundations of Modern Analysis. Pure and Applied Math., Vol. 10
Dieudonne, J. Bookseller: Zubal Books
(Cleveland, OH, U.S.A.) Price: US$ 9.49
Shipping within U.S.A.:US$ 6.50
Book Description: Academic 1960, 1960. 361 pp., hardback, ex library, else text and binding clean, tight and bright. Bookseller Inventory # ZB472982
 
Last edited by a moderator:
  • #39
notice there is a dearth of books listed for elementary diff eq since few of them inspire much admiration among people. on the other hand i have found some amazing bargains for you, including courant, apostol, hurewicz, hoffman/kunze, and dieudonne, at prices about 1/5 to 1/10 those often seen. sorry about kitchen. its a nice book but at that price it is absurd to buy it, given that copies of fleming, dieudonne, courant, etc... exist for so much less. almost anyone of these books will give you an enormous amount of education. i have also shortchanged complex analysis, but you will find another example on henry helson's website. he is a student of loomis i believe, and former berkeley professor who writes excellent books and publishes and sells them himself at reasonable rates, with some written by others. he has a linear algebra book too but i have not seen it.
 
Last edited:
  • #40
here are two more really good, really cheap books:

Elementary Theory of Analytic Functions of One or Several Complex Variables
Henri Cartan
Format: Paperback
Pub. Date: July*1995
B&N Price: $13.95
Member Price: $12.55
Usually ships within 2-3 days

also: Differential Forms
Henri Cartan
Format: Paperback
Pub. Date: July*2006
NEW FROM B&N
List Price: $12.95
B&N Price: $11.65*(Save*10%)
Member Price: $10.48
 
  • #41
For any mathematicians (pure or applied) did you guys intern anywhere during your summers? I am trying to find places where an applied mathematics major could go and intern during the summer (freshman). Maybe I could go abroad? Typically, does a math major do research over the summer or intern if he opting for a pHd? Does it have to be necessarily math related? Also, for an applied mathematician, what would you say is the most important area to know? Would it be ODE's / PDE's. I might be interested in going into quantitative finance, or something like biological math. This summer, I want to try to focus on learning a range of math rather than a depth of math (i.e. only studying Apostol, but not studying other areas of math like probability theory). Sure, I may not be a scholar in the end in any of the particular fields, but I can always go ahead and brush up later when the time calls for it (i.e. if I do a pHd). I find that the internet offers me the most versatility in learning different fields of math.
 
  • #42
I did not intern myself. Today there are several programs for math types in summer funded by VIGRE grants from NSF. Some schools also offer sumemr research opportunities but these are often voluntary activites by faculty, hence may fall short of volunteers. I.e. we are asked to do it for free, and that is something hard to sustain for long.
 
  • #43
hurkyl, i do not see how you can resist getting paid plus free tuition to study something interesting. how can you lose? it also adds to your resume for pay increases, new job opportunities, etc. i say grab it. you will do it easily. you are really strong mathematically. I am sure of this.
 
  • #44
Becoming a mathematician part 6) basic graduate books

Here are some foundational graduate books for future professionals. * means an especially high level recommended book.

Grad math books:

Algebra:
1. *Lang, Algebra,

2. Jacobson, Basic algebra 1,2.

3. Dummit - Foote, Algebra

4. Hungerford, Algebra

Reals

5. Measure and Integral: An Introduction to Real Analysis
Richard L. Wheeden, Antoni Zygmund

6. Royden, Real Analysis

7. Rudin, Real and complex analysis

8. * Functional Analysis, Riesz - Nagy

Complex

9. Ahlfors, Complex analyhsis

10. Conway, complex analysis

11. *Hille, Complex Analysis

12. Complex Analysis in One Variable, R. Narasimhan,

Topology
13. Fulton, Algebraic topology

14. *Spanier, algebraic topology

15. Hatcher, algebraic topology.

16. Vick, Homology theory.
 
  • #45
A remark for graduate students that they do not always seem to understand: Your instructor in a basic graduate course is often an expert in the field, at least on a level with many authors, although perhaps not all, of basic books. Hence it is not to be expected that the instructor will slavishly plow through a standard book on the topic, but may well merely present the material as best suits him or her. Do not be automatically disappointed if your instructor lectures from his/her own notes as they are often actually superior to what is found in many books. At the least the lecturer will probably select from the best presentations available for each topic.

This is a plus for the student. I am having difficulty citing here standard books for each subject, since at this level the presentation given in class is normally better than that found in anyone book, for one thing as it is more up to date, being given by a practicing professional. I.e. at this level the best instruction is often obtained in person rather than from books.
 
Last edited:
  • #46
Another remark: You will notice that all the books I have cited are theoretical ones, on specific bodies of theory, rather than being say problem books. This is the way I was taught, proving theorems. We were expected to find and work problems on our own.

But in Russia e.g., there is a wonderful tradition of problem solving and problem teaching. This type of activity was what brought me to math in high school but was slighted in my college instruction. Nonetheless it is gresat fun, and leads well toward the experience of doing a PhD and solving open problems.

Thus it would be good to list some books of problems, but I will have to do some research to find them.
 
  • #47
Hey mathwonk, would you mind doing a little comparison between pure math and applied math? As in the types of classes you'll take in each major, their differences, what you can do with each degree, etc

I am starting my undergraduate studies in September. While I don't have to decide on my major until I am done by first year, I'm still kind of curious as to how it all works.

Awesome thread, by the way. I've been reading it since you've started it.
 
  • #48
Well I really don't know anything about applied math, but i gather you should go heavy on the ode, partial de, and numerical analysis courses.

Thanks for the feedback.

I have been dominating the discussion but I want to explicitly solicit reports from other math people on their experiences in school, getting ready, what helped, what was a problem, what led to productive results at work, etc,...

Perhaps Matt could shed some light on his journey to a math PhD, and Hurkyl on his path to gainful employment, and J77 on his life as an applied math guy. Also physics guys like Zapper and others could help us with input on what math you really need if you might want to get into physics, or mathematical physics.

My friends in physics have emphasized group representations, but that was a long time ago. More recently it has been Riemann surfaces and algebraic geometry.
 
Last edited:
  • #49
My older son was a math major with numerical emphasis at Stanford, and now does web based internet stuff. He likes it. He also needs some business skill, as in a company you have to manage people who work for you, motivate, sell, service, hire and fire, and educate customers and clients.
 
Last edited:
  • #50
My wife was also a math major and is now a pediatrician. Math is not her main resource but all dosages require mathematics to scale them to suit each child by weight. I may be trivializing her math usage, but math majors can do a lot of things because they can reason and calculate well. She also needs to manage people and service customers.

Besides her ability to deal with all people she meets, her main skill that impresses me is her terrific diagnostic ability. She actually saves lives when she detects a serious infection by its outward signs. This is deductive ability appield to real life emergencies.
 
  • #51
One thing i can guarantee, everyone needs to take linear algebra, pure applied, whatever. The thing that is so frustrating about the AP courses in high school is their focus on calculus instead of linear algebra. I.e. linear algebra is easier than calculus, more important for more people than calculus, and even a prererquisite for understanding calculus.

So it sems odd to make calculus the focus of high school AP courses instead of linear algebra. Unfortunately no one listens to math professors when planning math education curricula.
 
Last edited:
  • #52
Some catch up points:

1. I didn't intern, some do, though, at investment banks and so on

2. I wouldn't bother with a research program with the aim of getting research under your belt if your intention is to do pure maths; it seems highly unlikely that anything you do will be representative of a pure PhD. However it can be a good experience of how mathematicians work, and you might get a glimpse of the future.

3. Books: I'd like to weigh in with some none analysis stuff, at the graduate level,

a) Fulton and Harris 'Representation Theory' for anyone considering doing algebra or theoretical physics. (contains all you need to know about semi simple lie algebras)b) James and Liebeck 'Representations and Characters of Groups' (brilliantly written intro to complex reps of groups)

c) LeVeque 'fundamentals of number theory' (all the basics)

d) Cox 'primes of the form x+ny' (very good intro to things like class field theory, must know number theory first, eg quadratic reciprocity first)

e) Weibel 'Introduction to Homological algebra' (all you wanted to know about homology theory but were afraid to ask, even introduces derived categories which are indispensable these days)

f) Alperin 'Local representation theory' (this is very specialized, but very accessible, worth a look for the out and out group theorist)My reasons for holding back are that I have a background in the UK and it is completely unrelated to the story unfolding here: there is no such thing as major and minor for a degree, you pick the subject whilst you're in high school that you'll do in university, and do it from day one when there. Doing a PhD in the UK is also vastly different: all those things that are taught in a US program are either things you're expected to know before you start or things you're expected to teach yourself if they're relevant to your area.

If you want to do a PhD in the UK, I'd very strongly recommend going to Cambridge to do Part III first (and this applies to international students too; I know plenty of Americans who did that year before going back to the states for their PhDs). It is the mathematical equivalent of basic training in the marines.

Don't let your mind go fallow either (one reason I've been posting here frequently in the last week is because I've got mathematicians block, and I'm trying to keep my mind active until it pops back into doing my research) and don't be afraid to look outside your area of interest. I see too many people dismiss something as being 'rubbish' just because it marginally falls outside their narrow ideas of what maths ought to be. It is to the UK's discredit that right now people are graduating with PhDs in this country in maths yet they don't know what a Riemann surface is, they've never seen any category theory, don't know a single cohomology theory. I can forgive any mathematician for not knowing what a sheaf is, but not for being ignorant of Galois theory, yet even that is missing from many of their memory banks.
 
Last edited:
  • #53
matt grime said:
If you want to do a PhD in the UK, I'd very strongly recommend going to Cambridge to do Part III first (and this applies to international students too; I know plenty of Americans who did that year before going back to the states for their PhDs). It is the mathematical equivalent of basic training in the marines.
I would say that other masters courses - particulary those by advanced study and research - are equally beneficial as preparation for a PhD in the UK.

And with the Part III, you can also specilise in Applied or Pure, right? So what's so special about the Part III?

I like how this thread's going, and that question's not a swipe at Cambridge, I'm interested...
 
  • #54
" And with the Part III, you can also specilise in Applied or Pure, right?"

No, you get to do whatever the hell you like.

In Part III the lectures are intense, far reaching, there are many different courses, far more than the average university is capable of handling, and widely recognised at international level to be outstanding. None of that applies to other taught masters courses in the UK, which then to be very narrowly focused on one particular area. You want to do graduate level courses in QFT, Lie Algebras, Differential Geometry, Non-linear dynamics and Galois Cohomology of number fields? Could be arranged, depending on the year (that was a selection of courses available when I did it). Where else would you be able to do that?

Feel like finding out about modular representation theory, combinatorics, functional analysis, fluid mechanics, and numerical analysis? Again, quite likely you can do that.

Of course, why you would want to do that is a something else entirely, but in terms of scope of work and expectations placed upon you it is the best preparation out there, far more so than most (ifnot any, but I can't bring myself to make such sweeping statements) MSc's by research, and certainly more so than any MMath course.

If you even want to do a PhD in maths at Cambridge, they will demand part III, and many other places use it as a training ground and ask their students to go there.

The reason it is the best is because in some sense it is 'the only': there is no other university with the resources to be able to offer a program like it. Even Oxford can't compete, and most UK maths departments are just too small to offer anything comparable.
 
  • #55
Matt's remarks on differences in expectations in US, UK remind me of a talk I heard at a conference. The speaker said something like, "this proof uses only mathematics that any sophomore undergraduate would know", then paused and added, "or here in the US, maybe any graduate student". This is true and getting worse.

Not only do we need to teach incoming grad student essentially beginning abstract algebra and analysis, but increasingly today also advanced calculus, and even basic proof writing in some cases.

This all goes back to the same problem - almost non existent training in basic math in high school, because of the ill conceived AP program. In the 1960';s there was a very ambitious and excellent set of hiogh school level books put out by the SMSG (school mathematicsa study group) via Yale University Press.

These constitute an excellent high school math preparation, including linear algebra, geometry, and calculus, but they are very hard to find now, being long out of print. One possible place to find them is math ed libraries in colleges of education.
 
Last edited:
  • #56
Thanks for the book list Matt. That would be my next category, specialized books, as opposed to the ones so far covering what "everyone" should know. It is harder to choose them though, so I appreciate the help from various perspectives. Many will disagree already with some of my choices for "everyone". A future algebraic geometer should ideally know at least abstract algebra, commutative algebra, homological algebra and sheaves, complex analysis (one and several variables), algebraic topology, and differentiable manifolds, hopefully differential geometry. Also something about projective geometry and plane curves. Not to know cohomology and some Hodge theory also seems unthinkable. I myself have alkways felt handicapped by a lack of knowledge of group representations.

Real analysis is used less, but it is used to prove Serre's duality theorem in cohomology. It is hard to prove finite dimensionality of a priori infinite dimensional cohomology spaces without some functional analysis.

But I want to be careful here of blowing away young students by overwhelming them with "what one should know" as the list is potentially unbounded. Getting a PhD means having basic training, then acquiring a lot of knowledge about something very restricted, and verifying some little new insight about it. All the rest of this information just builds up over time. I found it very important after my first post PhD job, to have a learning seminar every week, reading some good paper, or even teaching basic information to each other. You would be amazed how many PhD professors will come up quietly and say, I have always wanted to know what a differential form is, or a manifold, or a sheaf, or a toric variety, or algebraic variety, or chern class,... I myself taught galois theory and wrote my "graduate" algebra notes the way i did because i had never understood that subject. Just keep pecking away, learning, teaching, and doing mathematics, and eventually you will know a lot.
 
  • #57
As to my own history of things I did and didn't know, I entered college knowing what a group is, at least the definition, but upon entering grad school still did not know what an ideal was. Hence the note on the bulletin board with the "pre class" reading assignment for first yr algebgra, chapter 4 of Zariski - Samuel, was energizing. This chapter covered noetherian rings, including Noether Lasker decomposition theory of ideals, prime, primary, principal ideals, unique factorization, localization, etc etc..

Nontheless, after one year I was among the subset who passed algebra quals, but not my friend who had written his senior undergrad thesis on Zariski's main theorem in algebraic geometry. Tests are so odd, but I did not complain, having usually been treated better than I derserved by them.

There is a big difference between "knowing" a topic and understanding it too. Having read Lang's Analysis II, I "knew" the implicit function theorem to be a special case of the rank theorem, that a smooth function with locally constant rank near p, can be written locally, after a smooth change of coordinates, as a composition of a linear projection and a linear injection.
But when a professor remarked that the theorem meant you "can solve for some of the variables in terms of the others", I thought "huh?".

Another time at a high level research and instructional conference in algebraic geometry I sat next to Professor Swinnerton - Dyer and his student. The student was amazed that the Fields medalist speaker was hesitating over some elementary point in projective geometry, and his Professor was assuring him "they only teach that at Cambridge nowadays".

Perhaps it would help if Matt would describe a pre college and college preparation in the UK for maths. That could give us a more ideal pattern than ours here. British books are often the best to learn from as well, compared to American ones, since they seem to be written more often by people who know the language.
 
Last edited:
  • #58
I might mention, in case it is relevant for someone, I have not tried to talk to the upper upper percentile of math persons. I have tried to keep it reasonable for a bright math loving student. There are a few people who can handle anything thrown at them, just not most of us.

From 1960-1964 there were undergrads I knew at Harvard, maybe even the typical very good math major, who took the following type of preparation: 1st year: Spivak calculus course, plus more; second year: Loomis and Sternberg Advanced calculus, Birkhoff and Maclane, or Artin Algebra; 3rd yr: Ahlfors and maybe Rudin Reals and Complex; 4th year: Lang Algebra, and Spanier Algebraic Topology.

(Actually most of those courses did not even use books, just the professors notes, but those books are an approximation.)

Others actually began as freshmen in graduate courses. (I myself was briefly placed in an advanced graduate course in mathematical logic by Willard Van Orman Quine, but I did not care for the pace of it.) These were students who went to Princeton or Berkeley or Harvard afterwards for PhD.

Many added more advanced topics courses to these and wrote a thesis. One kid provided a small but key step for the classification of finite groups.

Indeed I might have even survived this regimen with much better study skills. But there were also people like Spencer Bloch, and John Mather there as undergrads who are famous figures in mathematics now. And the program there was ideal for them.

The point is not to compare ourselves too much with others, just to use them as inspiration, not discouragement, and keep on at our own pace, enjoying it as much as possible.
 
Last edited:
  • #59
Mathwonk, is there any chance that I could PM you my Email, and you send me a typical syllabus for what you teach in Calculus I and II and with what book? For me, Calculus I and II were taught out of Stewart's.

I'm currently in calculus II, and sometimes I think that my calculus I course was a bit "mickey mouse"-- as I never really had to study beyond doing the homework, and I found the exams relatively easy compared to what I was expecting. I honestly don't like the idea of having an easy course, and I want to excel when I eventually reach graduate school--not be blown away by people that had more rigorous courses during their undergrad.

Although I'm a physics major, I'll admit that my favorite part of physics is the math.
 
  • #60
mathwonk said:
My brother was Bill Monroe's fiddler in college.

Wow!

By the way, to my knowledge, the only mathematicians posting regularly on this site are Matt Grime and me.

Shmoe (post #20 in this thread) is a fairly regular poster.

mathwonk said:
Also physics guys like Zapper and others could help us with input on what math you really need if you might want to get into physics, or mathematical physics.

mathwonk said:
Today especially it is important to know some physics even if if you only plan to do math.

And vice versa, i.e., today especially it is important to know some math even if if you only plan to do physics. In this, I including some pure mathematics, otherwise you might end with nonsense like 0 = 1, as I pointed out in https://www.physicsforums.com/showthread.php?t=122063".

When I was a student I enjoyed mathematics courses both because I was interested in the applications of mathematics to modern theoretical physics, and because I enjoyed mathematics. Here are some of the math courses that I took.

Real Analysis: Analysis in Euclidean Space by Kenneth Hoffman
Measure Theory: The Elements of Integration by Bartle
Algebra: Basic Algebra I by Jacobson
Topology: Topology a First Course by Munkres
(Baby) Functional Analysis: Introductory Functional Analysis with Applications by Kreyszig
Representation Theory: Linear Representations of Groups by Vinberg

Below, I repeat a post that I wrote in response to a question about math references for physicists. The topics are pretty basic for mathematicians, but many physics students never see *any* of this stuff done in a "mathematical" style, and all the books listed below, except maybe Nakahara, are written in this style.

It is probably impossible for anyone person to learn all the mathematics useful for physics, so you have choose what mathematics you want to study and how much time you want to spend on it.

Typically, mathematical physics courses emphasize techniques for solving differential equations, e.g., special functions, series solutions, Green's functions, etc. These techniques are still very important, but, over the last several decades, abstract mathematical structures have come to play an increasingly important role in fundamental theoretical physics. Consequentlly, useful courses include real/functional analysis, topology, differential geometry (from a modern perspective), abstract algebra, representation theory, etc., and, usually, should be taken from a math department, not a physics department.

These courses, supply vital background mathematics, and, just as importantly, facilitate a new way of thinking about mathematics that complements (but does not replace) the way one thinks about mathematics in traditional mathematical physics courses.

A number of good books on "modern" mathematics exist. Among these, my favourite is https://www.amazon.com/gp/sitbv3/reader/104-8106425-5831130?%5Fencoding=UTF8&asin=0226288625#reader-link"&tag=pfamazon01-20 by Robert Geroch. Geroch purposely and provocatively chose his title to indicate that, these days, mathematical physics includes topics other than those covered in more traditional mathematical physics courses. He starts with a few pages on category theory!

Geroch's book contains a broad survey of abstract algebra, topology, and functional analysis, and it does a wonderful job at motivating (mathematically) mathematical definitions and constructions. Surprisingly, since Geroch is an expert, it contains no differential geometry. Also, its layout is abominable.

At slightly lower levels are https://www.amazon.com/gp/sitbv3/reader/104-8106425-5831130?%5Fencoding=UTF8&asin=0521829607#reader-link"&tag=pfamazon01-20 by Chris Isham.

Geroch's book should be supplemented by more in-depth treatments of topics. For example, a good mathematical introduction to group theory is https://www.amazon.com/gp/sitbv3/reader/104-8106425-5831130?%5Fencoding=UTF8&asin=0521248701#reader-link"&tag=pfamazon01-20 by Shlomo Sternberg.

Also, none of the surveys that I listed treat fibre bundles, which are so important in modern gauge theories, and in other areas. Treatments include https://www.amazon.com/gp/product/9810220340/?tag=pfamazon01-20 by Chris Isham.

This is just the tip of the iceberg - there are many, many other good books including Nakahara, Choquet-Bruhat et al., Reed and Simon, Fulton and Harris, Naber, ...
 
Last edited by a moderator:
  • #61
Boy that's a great post, George. Thank you. And your undergraduate basic training is very top notch mathematically.

Beeza, I usually teach calculus myself out of Stewart or Edwards and Penney, or such a standard good book. I like Stewart. It depends on how the course is taught. Most of my students are not math majors, and I often give them what may be weaker than a strong AP high school course.

I try to give whatever is appropriate to the audience and the course level and most of the time I am teaching non honors basic plug and chug calc.

But I always try to put some theory in there. Like when I do exponential functions, I really try to explain why they work the way they do. I will encourage kids to learn to derive the homomorphism law exp(a+b) =exp(a)exp(b), from the differential equation exp' = exp, and exp(0) = 1.

or use the differential equation ln' = 1/x and ln(1) = 0, to prove ln(ab) = ln(a)+ln(b). And I often ask them to elarn to prove that a function with derivative zero is constant, from the mean value theorem.

most people teaching from stewart skip these proofs, or even skip the statement of the MVT and the IVT.

If I happen once in a blue moon to get an honors class, not a spivak class mind you, just a basic honors class, I do try to put in a little more. It is usually very hard for my scholars who are coming in often with a no proofs at all AP course, often into my second semester honors class.

E.g. last time i taught regular honors integral calc, I not only proved all monotone functions are riemann integrable, but also covered power series as special cases of series of functions convergent in the uniform norm.

I discussed the equivalent sum norm, euclidean norm, and max norms on R^n, and then treated the analogous norms on function spaces, L^1, L^2, and sup norm. We proved differentiability and integrability conditions for convergent series of functions and applied to show you can integrate and differentiate power series term by term.

I can't remember if I proved the inverse function theorem but I did prove that all continuous functions on a closed bounded interval assume a maximum value.

This is the sort of approach in Lang's book Analysis 1, listed above.

The idea is to put some experience with proofs into even the basic course, and show a little something I might enjoy myself in the honors course.

You can make a pretty strong theoretical course out of stewart or edwards and penney if you prove all the theorems in there. most of the proofs are in the appendix anyway. it is just sort of the attitude towards proof in those books that is missing.

and which problems do you work? edwards and penney responded to market pressure by adding thousands of easy problems at the beginning of their problem sets, but the harder more interesting ones are often still there in the back of the problem set.

some things though are gone. e.g. calculus was invented by Newton primarily to do physics, and the old editions of E-P had kepler's laws derived, but now some of those are left out.

to get what you missed just start reading courant or spivak, or courant and john, or apostol. you'll see the difference.
 
Last edited:
  • #62
honors freshman calc notes

Here is one chapter from my first semester freshman honors calc course notes. i have only taught this stuff that once. I wrote these notes up just for that class, and i do it differently every time. In particular I had never thought of the local boundedness approach before.

Here is one chapter from my first semester freshman honors calc course notes. i have only taught this stuff that once. I wrote these notes up just for that class, and i do it differently every time. In particular I had never thought of the local boundedness approach before.

2300H. Chapter Seven. Bounded and Unbounded Functions
We want to discuss the basic notions of boundedness and unboundedness for functions, which just means whether their values can become arbitraily large in absolute value on their domains or whether there is some absolute value that is never exceeded. Later on in the case of functions that are also differentiable, we will use derivatives to actually find points where some functions take on maximum and minimum values. But not all functions actually assume such maxima or minima, not even all bounded functions. Thus it is useful to know a condition that guarantees a function has a maximum value in a given domain. The most fundamental condition says that if a function is continuous on a closed and bounded interval, then the function is bounded there (above and below) and it takes on a maximum value and also a minimum value.

Notation: "iff" = "if and only if".

Definition of boundedness:
Let f be a real valued function defined on a set S. Then we say f is "bounded on S", (or simply "bounded" if S is understood), if all of the values of f on S are bounded by some fixed number, in absolute value. More precisely, f is "bounded on S" if and only if there is some real number M >= 0 such that for every x in S, |f(x)| <= M.
In symbols, f is bounded on S if and only if for some M>=0, and all x in S, |f(x)| <= M.
Note this guarantees that f is bounded both above and below, since then, for every x in S, we have - M <= f(x) <= M.

Remark: If S is a finite set, every function defined on S is bounded on S. Just take the bound M to be the largest of the finite colection of numbers {|f(x)|, for x in S}. Hence boundedness is only in question when the domain S of f is an infinite set, such as an interval of positive length.

Definition of unboundedness: A function f defined on S is unbounded on S, if it is not bounded on S, i.e. if no matter how large we take M to be, there is always some point x in S where |f(x)| > M. In symbols, f is unbounded on S if and only if for all M>=0,there is some x in S with |f(x)| > M.

Remark: Note how the quantifiers (for some, for all) in the unboundedness statement are the opposite of those in the boundedness statement. Moreover it is sufficient to say that |f(x)| can be made larger than any natural number n, i.e. f is unbounded on S if and only if
(for every integer n >= 0)there is an x in S): (|f(x)| > n).

Examples The function f(x) = x is continuous and unbounded on the whole real line, but is bounded on every finite interval. The function cos(x) is bounded and continuous on the whole real line and hence also bounded on every finite interval. On the other hand the function g(x) = 1/x is bounded on the infinite interval [1,infinity), and unbounded on the finite interval (0,1). It is no accident that g(x) is discontinuous at 0, and has no continuous extension to the closed interval [0,1]. Note also that g(x) is actually unbounded on every interval of form (0,1/n) for every n. The more interesting function h(x) = cos(x)/x is also unbounded on every such interval. (You should graph these functions and be familiar with them.) This phenomenon of being unbounded on every interval about a point, no matter how short the interval, is called “local” unboundedness.

More precisely,

Definition of local unboundedness: If f is defined on some deleted neighborhood D of a point a (i.e. some set of form (a-?, a) union (a, a+?), or on some open interval with a as endpoint (i.e. some set of form (a-?, a) or (a, a+?)), but not necessarily at a, then f is “locally unbounded at a” (or “locally unbounded near a”) if and only if f is unbounded on every deleted neighborhood of a. In symbols, f is “locally unbounded at a” if and only if no matter how large we take M to be, and no matter how small we take ? to be, there is some point x in (a-?,a+?) where f has a value larger than M in absolute value. In symbols, f is “locally unbounded at a” if and only if
(for all M >= 0)(for all ? > 0)(there exists x)such that (0 < |x-a| < ? and |f(x)| > M).

Definition of local boundedness: If f is defined on some deleted neighborhood D of a point a, or on some open interval with a as endpoint, then f is “locally bounded at a” (or “locally bounded near a”) if and only if f is bounded on some deleted neighborhood of a. Equivalently, there is some bound M and some deleted ? neighborhood of a where f is bounded by M, in absolute value. In symbols, f is “locally bounded at a” if and only if
(for some M >= 0)(for some ? > 0)(for all x)( 0 < |x-a| < ? implies |f(x)| <= M).
In case f is only defined say to the right of a, this would read
(for some M >= 0)(for some ? > 0)(for all x)( 0 < x-a < ? implies |f(x)| <= M).

Continuous functions are locally bounded. This is a basic property of continuous functions. More precisely,
Theorem: If f is continuous at a, then f is locally bounded at a.
proof: Take any positive number e> 0, (such as e = 1). then by definition of continuity there is some ? > 0 such that for all x with
|x-a| < ?, we have |f(x) - f(a)| < e. We claim then M = (|f(a)| + e) is a bound for f on the interval (a-?,a+?). I.e. for all x with |x-a| < ?, we have |f(x) - f(a)| < e, but in general |A| - |B| <= |A-B|, so for |x-a| < ?, we have |f(x)| - |f(a)| <= |f(x) - f(a)| < e, thus |f(x)| - |f(a)| < e. Hence by adding |f(a)| to both sides of this inequality, we get |f(x)| < |f(a)| + e = M, for all x with |x-a| < ?. This proves f is locally bounded near a. QED.

Note this means that a function like f(x) = 1/x can be locally bounded at every point of (0,1), since it is continuous at every point of (0,1), and yet be unbounded globally on (0,1). This is only possible because the interval (0,1) is open. I.e. we have the following basic theorem:

Theorem: If f is locally bounded at every point of the closed bounded interval [a,b], then f is globally bounded on [a,b].
proof: We know f is locally bounded near a, so at least there is some x with a < x < b such that f is bounded on the interval [a,x]. We just want to show we can move x to the right all the way to b and still have f bounded. So consider the set S = { those points x in [a,b] such that f is bounded on [a,x] }. Then a belongs to S, but all elements of S are bounded above by b, so S has a least upper bound, say L. Then we can conclude that f is bounded on [a,x] for every x with a <= x < L, but f is not bounded on [a,x] for any x with L < x <= b.

We claim L = b. We will prove it by contradiction. Since a is in S and L is an upper bound for S, we see a <= L. Since b is an upper bound for S and L is the least upper bound, we see also that L <= b, so a <= L <= b. We claim L cannot be less than b. For if L < b, then since f is continuous at L, by assumption, f is locally bounded at L so there is some ? > 0 such that f is bounded by some M1 on (L-?,L+?) and by taking ? smaller if necessary, we will have L < L+? < b. Hence f is bounded by M1 on the interval [L- ?/2, L+ ?/2] and, since L - ?/2 is less than L, also f is bounded on the interval [a, L- ?/2] say by M2. Then if M = max{M1, M2}, f is bounded by M on the whole interval [a, L+ ?/2].

But this contradicts the fact that f is not bounded on [a,x] for any x with L < x <= b. I.e. L + ?/2 would be such an x. This contradiction shows that in fact L < b is impossible, so L = b.
Now we know that f is bounded on all intervals of form [a,x] with
x < b, and we claim that in fact f is bounded on [a,b]. Also f is locally bounded at b, say by N1, on some ? - neighborhood of b, i.e. on (b-?,b) for some ? > 0. On the other hand, since b - ?/2 < b = L, f is also bounded on [a, b - ?/2] say by N2.

Then f is bounded on [a,b) by max{N1, N2}. We have to include the endpoint b also but that is easy since it is only one point. I.e. if we take N = max{N1, N2, |f(b)|}, then f is bounded by N on [a,b]. QED.

Corollary: In particular, since every continuous function on [a,b] is locally bounded at every point, every continuous function f on [a,b] is globally bounded on [a,b].

From the corollary, we know the set of values of a continuous function f on [a,b] have an upper and a lower bound, and hence by the least upper bound axiom for real numbers, there is a least upper bound and a greatest lower bound for those values. We claim that for a continuous f on [a,b] that the lub of its values is actually a value, the “maximum value”, and the glb of the values is also a value, the “minimum value”. Thus a continuous function on a closed bounded interval assumes a maximum value and a minimum value on [a,b].

More precisely,
Theorem: If f is continuous on [a,b] and if M is the lub of its values on [a,b] while m is the glb of those values, then there is some point x0 in [a,b] where f(x0) = m, and some point x1 in [a,b] where f(x1) = M.
proof: (for M). Suppose to the contrary that f never takes the value M. Since M is the least number >= all the values of f on [a,b], then f takes values arbitrarily near M from below, i.e. for every n > 0, there is a value of f between M - 1/n and M. Thus for every n > 0, there is some x in [a,b] such that M - 1/n < f(x) < M. Thus the reciprocal function 1/(M-f(x)) is greater than 1/(1/n) = n at x. Thus if f never equals M, then the reciprocal function 1/(M-f) is both continuous and unbounded on [a,b]. This is a contradiction to the previous corollary. QED.

Corollary: If f is continuous on [a,b] then there are points x0 and x1 in [a,b] such that for every x in [a,b], we have f(x0) <= f(x) <= f(x1). We say f takes its maximum value at x1 and its minimum value at x0.
 
  • #63
honors integral calc

Here are some of my second semester honors calc notes for freshmen. sorry about the screwed up fonts. i hope it is readable somehow.

2310H Sequences and series
We want to use limit processes to extend our reach from the familiar to the unfamiliar, by approximating some exotic functions and numbers in terms of more familiar ones. E.g. we will approximate irrational numbers like e and <pi>, by simple rational numbers. And we will approximate exotic functions like ex, ln(1+x), sin(x), and arctan(x), by polynomials. Remember, to say some irrational number is a limit of rational numbers just means we can approximate the irrational number as closely as we like by rational numbers. Also to say a function is a limit of polynomials means we can approximate the given function as closely as we like by polynomials. For this we must decide what it means for two functions to be near each other. Does it mean all their values are uniformly near? I.e. that there is not much distance between their graphs?” Or does it mean there is not much area between their graphs? Or something else? We will discuss the various choices below.

Definition: If S is any set, a sequence with values in S is simply a function a:N-->S where N is the "natural numbers", i.e. the positive integers. We denote the value a(n) by an and display the whole function as its sequence of values: a1,a2,...,an,...

Remark: It is not essential that the indices begin with 1; they could begin with 0, or -4, or 1000. The important thing is that it begin somewhere and go on up to infinity. I.e. it is only infinite in one direction, upwards.
We often use letters to denote the values of a sequence that remind us of the nature of the elements of the set S.

Example 1: If S is the real numbers we might call the elements xn, and write the sequence as {xn} or x1,x2,x3,...,xn,...
Example 2: If the set S is the Euclidean plane R2, we might write {pn} or p1,p2,p3,...,pn,... for a sequence of points in the plane.
Example 3: If the set S is the set of continuous functions on the interval [a,b], we might write f1,f2,...,fn,... or {fn} for the sequence of functions.

In all three of our examples, we can add, subtract, and multiply our elements by real numbers. We want to define next a notion of "size" or "length" or "absolute value" for our elements.
ex.1: For real numbers define the absolute value of a number x to be |x| = its absolute value.
ex.2. For a point p =(x,y) in the plane, define its absolute value to be the Euclidean distance from the origin, |p| = sqrt(x^2+y^2).
ex.3. For a function f on [a,b] there are several natural choices, which yield different results. The one suited to our present purposes is called the "sup norm", which is the maximum of all the absolute values |f(x)|, i.e. ||f|| = the global maximum of the function |f|. Thus ||f|| is the maximum of the absolute values |f(x)| of f evaluated at every point x in [a,b].

Thus ||f|| is the height of the highest point of the graph of the function y = |f(x)|, over the interval [a,b]. We know from a big theorem in my 2300H notes, that there exists such a maximum.
These notions of length lead a notion of distance between two objects, and hence of a notion of "epsilon neighborhood" centered at one object:

ex.1: Given two real numbers x,y, their distance apart is |x-y|. For e > 0, the e - neighborhood of x, is the open interval (x-e, x+e) of all real numbers closer to x than e.


ex.2: Given two points in the plane p1 = (x1,y1), and p2 = (x2,y2), their distance apart is |p1-p2| = sqrt([x1-x2]^2 + [y1-y2]^2). Given e > 0, the e - neighborhood of p, is the open disc of radius e, centered at p, of all real points in the plane closer to p than e, in the usual "Euclidean norm".


Remark: We get the same notion of convergence in the plane, but not exactly the same notion of distance, by saying that the distance between two points p1 = (x1,y1), and p2 = (x2,y2), is the maximum of |x1-x2|, or |y1-y2|. I.e. by seeing how far apart their x and their y coordinates are, and taking the larger difference as the distance between the points.

Then given e > 0, the e - neighborhood of p, would be the open square of radius e, centered at p, all real points in the plane closer to p than e, in the "maximum norm".

There is a third natural notion of length and distance for points in the plane, called the “sum norm”, where the length of p = (x,y) = |x|+|y| is the sum of the absolute values of the coordinates. Then the distance from p1 = (x1,y1), to p2 = (x2,y2) is |x1-x2|+|y1-y2|, and the e - nbhd of p, is a “diamond” of radius e, centered at p:

ex.3: The three definitions of “length” we discussed in the plane all have generalizations to “size” of functions. The Euclidean norm generalizes to the "L2-norm" where a function has size = , the square root of the integral of its square. If we think of a function as a “vector” with an infinite number of components, this definition yields a related definition of “dot product” = which allows one to talk about the “angle” between two functions and perpendicularity of functions. This particularly use in approximating functions by sines and cosines, called the theory of “Fourier series”.
The sum norm generalizes to the integral of the absolute value. This was Matt's suggestion, and it is very useful in extending the notion of integrability of functions to more general functions than the ones Riemann’s definitions works for. Convergence using this notion of length, the “L1-norm”, leads to the theory of “Lebesgue integration”.

For our purpose of approximating functions by polynomials, it is useful to choose the generalization of the "max norm" we defined above. Thus the distance between two functions f,g in the max norm, is defined as ||f-g|| = maximum of all differences |f(x)-g(x)|, for all x in [a,b].
For given e>0, the resulting e - nbhd of f, is represented by a strip extending a distance e both above and below the graph of f. I.e. a function g is within a distance e of f if and only if its graph lies entirely in that strip.


Remarks: All our notions of length satisfy these basic properties:

(i) "triangle inequality" |a+b| <= |a| + |b|, |a-b| >= |a| - |b|.
(ii) “homogeneity”: |ca| = |c||a|, where c is a real number.
e.g. ||cf|| = |c| ||f||, for a function f and a constant c.
(iii) “non degeneracy”: |a| = 0 if and only if a = 0.

Although all three norms in the plane give the same notion of convergence, this is not true for their generalizations to functions. Here the sup norm is more restrictive than the L1 or L2 norms.
Exercise: If two continuous functions on [a,b] functions are everywhere within e of each other then their integrals are also within e(b-a) of each other hence also close. [Hint: Recall the monotonicity property of integrals, that f(x) <= g(x) for all x in [a,b], implies integral f <= integral g .]
In particular a function which is everywhere close to zero, has integral which is also close to zero. I.e. if a function is small in the sup norm, it is also small in the L1 norm. On the other hand a function can have integral very close to zero and yet can have some very large values. Hence a function can be small in the L1 norm and yet be very large in the sup norm. Here is one such: [imagine picture]

This function has sup norm equal to n, and yet has integral 1/(2n). So the sup norm approaches infinity while the L1 norm approaches zero. Thus convergence is different in these two norms.

Thus it is harder for functions to approximate other functions in the sup norm, which means that the limit function will retain more properties of the approximating functions. This suits us since we are interested in approximating very good functions like sin and exp, which have the same good properties of continuity and differentiability as the approximating functions we will use, the polynomials. (If on the other hand we wanted to define the notion of integral for functions with lots of discontinuities, we would use a norm like the integral norm which allows very continuous functions to approximate very discontinuous ones.)

Definition: A sequence {sn} in S, (where S is one of our three sets equipped with the appropriate distance), converges to an element s? of S, or simply {sn} --> s?, if and only if, for every e>0, there exists a positive integer N, such that whenever n >= N, then |sn-s?| < e. Note: to say |sn-s?| < e, is the same as saying s? - e < sn < s? + e. (Although we write a single absolute value here, in the case of functions this is the sup norm || ||.)

Note: To say {sn} --> s?, is equivalent to saying that {sn-s?}-->0.Remark: This means that no matter how small an e- neighborhood we describe around our limit point s?, after a certain element sN, all the rest of the sequence lies in that neighborhood. In particular if a sequence converges to s?, and we form a new sequence by throwing away the first billion elements of our old sequence, the new sequence also converges to s?. Thus whether or not a sequence converges, and what the limit is, is unaffected by any given finite number of elements of the sequence.
In particular, if a sequence converges to s?, then the new sequence formed by adding in a billion or so 1’s at the beginning of the sequence, still converges to the same limit. Thus there is no reason to expect to able to guess the limit of a sequence just by looking at the first hundred trillion elements or so.

Remark: Because all our notions of length satisfy the triangle inequality, it follows that the sum of two convergent sequences converges to the sum of the limits, and homogeneity implies that multiplying the elements of a sequence by a constant multiplies the limit by that constant. Non degeneracy implies that the limit of a sequence is unique, i.e. the same sequence cannot converge to two different limits. Of course these are intuitive properties we might expect. And they are indeed true. (You should prove them.)

To prove anything about existence of limits we need an axiom guaranteeing the existence of lots of real numbers. A surprisingly simple one suffices.
Completeness axiom: A non empty set of real numbers which has an upper bound has a least upper bound. I.e. if some number is >= than all numbers in the non empty set S, then there is some smallest number which is still >= all numbers in S.

Corollary: Every set of real numbers with a lower bound has a greatest lower bound.
proof: (Use minus the least upper bound of the negatives of these numbers.)

Remark: Since we said the real numbers are represented by infinite decimals, we can prove the completeness axiom as a theorem. E.g. given any collection of positive infinite decimals, none of which end in all 9’s, and which are bounded above say by N, choose as integer part the largest integer occurring as integer part of one of them. Then among all those reals having exactly that largest integer part, choose the largest tenths digit that occurs among these. Then among all reals in the set having exactly that integer part and that tenths digit, choose the largest hundredths digit that occurs. Continue... and you will construct a decimal that is the smallest number not smaller than any of your decimals. (Note this construction can give a decimal that does end in all 9’s, in which case you can choose a different representative which does not do so.)

Application: The sequence of positive integers {n} is not bounded above.
proof: If it were, there would be a smallest upper bound K. Then K-1 is smaller so K-1 is not an upper bound for all positive integers, so there is some positive integer N with N > K-1. But then K+1 > N, contradicting N being an upper bound for all positive integers. QED.

Corollary: The sequence {1/n} of reciprocals of all positive integers n, converges to 0.
proof: Given e > 0, choose N > 1/e. This is possible since the positive integers have no upper bound. Then for all n >= N, also n > 1/e. I.e. then 1/n < e. So for all n >= N, we have |1/n - 0| < e. QED.

Remark: With our convention that reals are decimals, the past two properties are also somewhat obvious by reasoning with decimals.

Corollary: If 0< r < 1, the sequence {r^n} converges to 0.
proof: Let s = 1/r > 1, and then given e > 0, choose N > 1/e(s-1), hence N(s-1) > 1/e. Then n >= N implies s^n >= s^N = (1+(s-1))^N >= 1+N(s-1) [binomial theorem] >= 1+(1/e) = (e+1)/e. Then 1/s^n = r^n <= e/(e+1) < e. QED.
another proof: given a>0 we want to find N so that n>=N implies that r^n < a. taking logs this is equivalent to n ln(r) < ln(a), i.e. since 0<a<1 implies that ln(a) < 0, this is equivalent to n > ln(a)/ln(r). Since the integers are unbounded above, just choose N > ln(a)/ln(r). Then n >= N implies also n > ln(a)/ln(r). Hence n ln(r) < ln(a), so after exponentiating, we get r^n < a, as desired. QED. Infinite series
Next we discuss “infinite sums” i.e. “convergent series”. Let {an} be any infinite sequence, and form another sequence of “partial sums” of the original sequence: s1 = a1, s2 = a1+a2, s3 = a1+a2+a3,..,
sn= a1+a2+...+an,...

Definition: We say "summation ai converges to a?", or “ = a?”, or “a? is the sum of the series summation ai”, if and only if the sequence {sn} of partial sums converges to a?, if and only if, for every e > 0, there is a positive integer N, such that, for all n>=N, we have |(sum from i = 1 to i = n of ai) - a?| < e.

Example: geometric series: If a is any real number and r is a real number with |r| < 1, then the series summation of ar^i from i=0 to i=n, converges to a/(1-r).
proof: By multiplying out the denominators, one checks that
= a/(1-r) - ar^(n+1)/(1-r), so |a/(1-r) - ?? | = |ar^(n+1)/(1-r)|. Since we know that |r|^n-->0, it follows that |ar^(n+1)/(1-r)| =
|ar/(1-r)| |r|^n -->0.
QED.
 
Last edited:
  • #64
more second semester honors calc notes

a bit more secomd semester homnors calc notes. these were completed by series for e^x, cos(x), sin(x), and proof of differentiability of convergent power series term by term.

Series of functions:
Example: power series:
Consider the functions x^n for n >= 0, and the formal geometric series expansion 1/(1+x) = 1 - x + x^2 - x^3 + x^4 - + ... We know the rhs equals the lhs for any choice of x with |x|<1 by the previous example. We claim this series of functions converges to the function 1/(1+x) on the lhs in the sup norm, on any interval [-r,r] where 0<r<1, (but not on all of
(-1,1)). I.e. since the partial sum 1 - x + x^2 - x^3 + ...+(-1)^n x^n =
[1/(1+x) - (-1)^(n+1)x^(n+1)/(1+x)], we have again that
| 1/(1+x) - (1 - x + x^2 - x^3 + ...+(-1)^n x^n)| = |x^(n+1)/(1+x)| for any real number x. Now since on the interval [-r,r] we have |x| <= r, and 1+x >= 1-r, it follows that for all x in that interval, ||x^(n+1)/(1+x)|| <= r^(n+1)/(1-r). Hence to show that || 1/(1+x) - (1 - x + x^2 - x^3 + ...+(-1)^n x^n)|| = ||x^(n+1)/(1+x)|| approaches zero, it suffices to show that r^(n+1) -->0, which we have done above. Thus 1/(1+x) = 1 - x + x^2 - x^3 + x^4 - + ..., for all x with |x|<1, and convergence holds in the sup norm on any closed bounded interval strictly contained in (-1,1). QED.

Exercise: (i) If a sequence of functions {fn} converges to f in the sup norm on [a,b], then the integrals also converge, i.e. the sequence of real numbers { } (integral of fn from a to b) converges to the real number (integral of f from a to b).
(ii) In fact the indefinite integrals Gn = , (integral of fn from a to x) which are functions on [a,b], also converge to the function G = (integral of f from a to x), in the sup norm.

Approximation of transcendental functions by polynomials
Example: ln(1+x):
By the previous example, 1/(1+x) = 1 - x + x^2 - x^3 + x^4 - + ..., for all x with |x|<1, and convergence holds in the sup norm on any closed bounded interval strictly contained in (-1,1). Consequently, by an exercise above, on any interval [-r,r] with 0<r<1, the series of indefinite integrals (starting at 0) of the series 1 - x + x^2 - x^3 + x^4 - + ..., converges to the indefinite integral of 1/(1+x).
I.e. the series x - x^2/2 + x^3/3 - x^4/4 + x^5/5 - x^6/6 ±... converges in the sup norm on [-r,r], to (integral of 1/(1+t) from t=0 to t = x)= ln(1+x). Thus ln(1+x) =
x - x^2/2 + x^3/3 - x^4/4 ±..., for each x with |x| < 1, and convergence holds in the sup norm on any [-r,r] with 0<r<1. Now because the series has alternating signs, it can be shown that it also converges for x = 1, to ln(2), and yields the amazing formula ln(2) = 1 - 1/2 + 1/3 - 1/4 + 1/5 - + ...

Example: arctan(x): The geometric series 1 - x^2 + x^4 - x^6 + - ..., converges to 1/(1+x^2), for each x with |x| < 1, and in the sup norm on any interval [-r,r] with 0<r<1. Hence the series of indefinite integrals, starting from 0, converges to the indefinite integral of the limit.
I.e. x - x^3/3 + x^5/5 - x^7/7 ±... = =(integral of 1/(1+t^2) from t=0 to t=x) = arctan(x), again in the sup norm on any closed interval strictly contained in the interval (-1,1). Again convergence holds also for x = 1, yielding the even more amazing
formula: <pi>/4 = 1 - 1/3 + 1/5 - 1/7 + 1/9 - + ...
Next we want to find series expressions for e^x, sin(x), and cos(x). Since the derivatives of these functions are no simpler than the functions themselves, we cannot proceed in the same way as before. We need some criteria guaranteeing convergence of sequences and series when we do not know what the limits are precisely. They all involve exploiting the notion of boundedness.

Convergence of monotone sequences
Lemma: A convergent sequence must be bounded. I.e. if {sn} converges, then there is some positive number K such that for all n, |sn| <= K.
proof: By definition of convergence, if {sn} converges to s, then given say e = 1, there is an N such that all elements after sN are within a distance 1 of s, so that for all n >=N, we have |sn| <= |s| + 1. Hence if we let K be the maximum of the numbers |s1|, |s2|,...,|sN-1|, |s|+1, then for all n, we have |sn| <= K. QED.

Remark: The converse does not hold, since the sequence
1,-1,1,-1,1,-1,... is bounded but not convergent.

There is however a class of bounded sequences of real numbers which does always converge, namely bounded monotone sequences.
Lemma: A bounded monotone sequence of real numbers converges.
Proof: If the sequence {sn} is bounded and monotone, say monotone increasing, let K be the least upper bound of the sequence. I.e. let K be the smallest number such that for all n, we have sn <= K. We claim the sequence converges to K. Let e>0 be given. Since K is the smallest number which is >= all elements of the sequence, the number K-e must be less than some element of the sequence. Suppose sN > K-e. Then for all n>=N, we have sN <= sn, by monotonicity. Since K is an upper bound of the entire sequence we also have K-e < sN <= sn <= K < K+e, for all n >= N. I.e.
|sn-K| < e, for all n >= N. QED.

Note: This gives a way to tell a sequence is convergent without explicitly finding the limit. Just find any upper bound for a weakly increasing sequence and you know the sequence converges even if you cannot determine what is the least upper bound, i.e. the limit. Similarly, if there is a lower bound for a weakly decreasing sequence, then that's equence also converges to its greatest lower bound.

Here is the analog for series, of convergence of monotone sequences.
Theorem: If {an} is any sequence of non negative numbers, the series (summation of ai from i =1 to i = infinity) converges if and only if the partial sums are bounded, i.e. if and only if there is some number K such that for all n, (summation of ai from i =1 to i = n)<= K.
proof: trivial exercise.

This leads to the following so called “comparison tests”.
Theorem: If (summation of ai from i =1 to i = infinity) and (summation of bi from i =1 to i = infinity) are two series of non negative real numbers, and if ai <= bi for all i, then the convergence of (summation of bi) implies the convergence of (summation of ai), and hence the non convergence of (summation of ai) implies the non convergence of (summation of bi).
proof: This follows from an earlier result because when the partial sums of one positive series are bounded, so are those of a smaller positive series. QED.

The idea that monotone sequences converge generalizes as follows.
Cauchy’s criterion and its applications
Definition: A sequence {sn} is called Cauchy, if and only if for every
e > 0, there is some N, such that, for all n,m >= N, we have |sn-sm| < e.
Exercise: Any convergent sequence is Cauchy.

Remark: In our three examples, the converse holds: in the real numbers, the plane, and the space of continuous functions on [a,b] with sup norm, every Cauchy sequence converges to an element of the same space.

Digression: Intuitively, to say a sequence is Cauchy, means the elements of the sequence are bunching up, but they might not converge unless there actually is a point of our space at the place where they are bunching. E.g., if our space were the real numbers, except zero had been removed, then the sequence {1/n} would still be Cauchy, but would not converge simply because we had removed the limit point. Since lots of sequences of rational numbers have irrational limits, Cauchy sequences of rationals do not always converge in the space of rationals. E.g. the sequence 3, 3.1, 3.14, 3.141, 3.1415, 1.14159,... of rationals, which converges to <pi>, (if the decimals are chosen appropriately), would be Cauchy in the rationals, but would not converge in the space of rationals. I.e. some spaces have “holes” in them, and a sequence could head towards a hole in the space and be Cauchy, but not have a limit in the space, just because the limit is missing from the space.
[There is a way to fill the holes in any space, i.e. a space with a distance can be enlarged so all Cauchy sequences do converge, by adding in a limit for every Cauchy sequence. This is one way to construct the reals from the rationals. Starting from any space with a length, consider the space of all Cauchy sequences in that space, and identify two Cauchy sequences {xn} and {yn} if the sequence of numbers {|xn-yn|} converges to zero. For instance the real number <pi> is identified with the Cauchy sequence 3, 3.1, 3.14, 3.141, 3.1415,... of rationals. Decimals give a very efficient way of picking usually one Cauchy sequence of rationals for each real number. Still the Cauchy sequences of decimals 1, 1.0, 1.00, 1.000, ... and .9, .99, .999, .9999, ... both represent the same real number.]

None of our 3 example spaces have holes, by the next theorem.

Big Theorem: In all three of our examples, every Cauchy sequence {si} converges to some limit in the given space.
proof:
Example (i) We do the case of real numbers first: define for each n, an = the greatest lower bound of the elements si in the sequence such that i >= n. Define bn = least upper bound of those elements si with i >= n. Then {an} is a weakly increasing sequence and {bn} is a weakly decreasing sequence, so both sequences {an} and {bn} converge by the previous corollary. Now the Cauchyness of the sequence {si} implies that |an-bn| converges to zero. Thus in fact both sequences {an} and {bn} converge to the same limit K. Then since for each n, all sk with k >= n, lie between an and bn, K is also the limit of the sequence {si}.

Here is another cute proof; we claim first (i) that every sequence has a a monotone subsequence, and then (ii) that every Cauchy which has a convergent subsequence, also converges itself.
proof of (i) Call a point sN of a sequence a “peak point” if all later members of the sequence are no smaller. I.e. sN is a peak point iff for all n >= N, we have sn >= sN. Now there are two cases: either there are an infinite number of peak points or only a finite number of them, maybe zero. If there are an infinite number of peak points, then the subsequence of peak points is weakly monotone increasing and we are done. If there are only finitely many peak points, then after the last peak point say sN, no element is a peak point. So every element sn with n >= N, has the property that there is a later element which is smaller. This allows us to choose a weakly decreasing subsequence. I.e. start from sN+1. then there is some sn with n >= N+1 and such that sn < sN+1. Let that sn be thes econd element of the subsequence. Then there is some later element sm such that sm < sn. Let that sm be the third element of the subsequence. Continue in this way.
proof of (ii) If a Cauchy sequence {sn} has a convergent subsequence {tm}, [i.e. each tm is one of the sn, and the t’s occur in the same order in which they occur in the original sequence], then the originals equence {sn} converges to the same limit as the subsequence.

To be precise:
Definition: Recall a sequence of reals is a function s:N-->R, where N is the set of positive integers and R is the set of real numbers. A subsequence is a function t:N-->N-->R which is a composition of a strictly increasing function N-->N, with the function s:N-->R. We some times write the element tm as s(n(m)), where we think of n as a function of m. Here by hypothesis, n(m) >= m, and also n(m+1) > n(m).

Ok, assume tm = s(n(m)) -->L. If {sn} is Cauchy we claim {sn}-->L also. So we just try to plod through the motions. I.e. let e>) be given. We must find N such that n>=N implies that |sn - L| < e. Ah my brain is waking up. OK, we know we can make all the later t’s close to L, by hypothesis that the sequence of t’s converegs to L. We also know that by the Cauchy hypothesis, we make all the later s’s close to each other. Since some of thos s’s are t’s, that should make all the alter s’s close to L too. OK, choose K so large that n >=K implies |tn-L| < e/2. And then choose M so large that n,m>=M implies that |sn-sm|<e/2. Then let N be the alrger of the two integers K,M. Then the element tN = s(n(N)) where n(N) >= N. So this implies that |tN-L| < e/2. Now let n >= N and look at |sn - L|. Since n(N)>=N, we know that |sn -s(n(N))| < e/2. Now we have
|sn - L| = |sn - s(n(N))+s(n(N)) -L| <= |sn - s(n(N))| + |s(n(N)) -L|
< e/2 + e/2 = e. That doos it I hope. QED.

Example (ii) For a Cauchy sequence of points {pn} = {(xn,yn)}, in the plane, both sequences {xn} and {yn} are Cauchy sequences of real numbers, since |pn| >= |xn|, |yn|. Hence {xn} converges to some x, and {yn} converges to some y, and then {pn} converges to (x,y).

Example (iii) If {fn} is a Cauchy sequence of functions on [a,b], then for each x in [a,b], the definition of the sup norm, forces the sequence of real numbers {fn(x)} to be Cauchy, hence convergent to some number we call f(x). This defines a function f, which we claim is continuous, and is the limit of the sequence {fn}.
To see convergence, let e>0 be given. We must find N such that for all n>=N, we have ||f-fn| < e. But we know the sequence {fn} is Cauchy in the sup norm, so for some N, we have ||fn-fm|| < e/3 for all n,m >= N. Since for all x, f(x) is the limit of the fn(x), it follows that for all x and all n >=N, we have |f(x)-fn(x))| <= 2e/3. I.e. given x, there is some m > N such that |fm(x)-f(x)| < e/3. Since for all n>=N, we have |fn(x)-fm(x)| < e/3, it follows that for all n >= N, |f(x)-fn(x)| <= |f(x)-fm(x)|+|fm(x)-fn(x)| < 2e/3. Thus for all x, and all n >= N, we have |f(x)-fn(x)| < e. I.e. {fn} converges to f in the sup norm.
Finally we claim the limit function f is continuous on [a,b], hence lies in the space we are working in. To prove this, let z be any point of [a,b]. To show f is continuous there, let e>0 be given and try to find d>0 such that for all x closer to z than d, we have |f(x)-f(z)| < e. This is a classic e/3 proof. I.e. choose N such that for all n,m >= N, we have ||fn-fm|| < e/3. Then we saw above that also for all n>=N, we have ||f(x)-fn(x)|| < e/3. Now fN is continuous by hypothesis, so there is a d>0 such that for all z closer to x than d, we have |fN(z)-fN(x)|<e/3. Then just note that
|f(z)-f(x)| = |f(z)-fN(z)+fN(z)-fN(x)+fN(x)-f(x)|
<= |f(z)-fN(z)| + |fN(z)-fN(x)| + |fN(x)-f(x)| < e/3 + e/3 + e/3 = e.
I.e. |f(z)-fN(z)| < e/3 because fN is closer than e/3 to f at every point of [a,b]. And |fN(x)-f(x)| < e/3 for the same reason. Then |fN(z)-fN(x)| < e/3 because fN is continuous at x, and d was chosen to make this true for fN since |z-x| < d. QED.
 
Last edited:
  • #65
Q: How familiar is this stuff to students? how many courses treat completeness of real numbers, cauchy convergence, boundedness of continuous functions, differentiability of series, with proofs, in high school AP? in college calc? in college honors calc?
 
  • #66
we don't cover any of this. I only saw most of the material in Apostol and Courant.
 
  • #67
well that's my point. that's the difference between a high school AP course and my honors college course. so if you are a good student, your AP course prepares you to begin my honors course in first semester. But not everyone teaches this way. In fact many people probably think i am a nutcase for teaching this stuff to freshmen, but i think you can do anything if you do it well and carefully.

this is some of the stuff Tate covered in our freshman honors calc course at harvard, but he also covered linear algebra, inner products and hilbert spaces, including the complex case, and differential equations.:biggrin:
 
Last edited:
  • #68
beeza, i recommend henry helson's honors calculus book, pretty cheap ($24 new, including shipping), from his website, and very well done for a short expert treament of high level calculus, by a retired berkeley professor.
 
Last edited:
  • #69
lets do some exercises. here is a little tiny proof: show that if f is a function which is bounded away from zero, i.e. f never takes values in some interval (-a,a) , then 1/f is bounded above and below.
 
  • #70
mathwonk said:
Q: How familiar is this stuff to students? how many courses treat completeness of real numbers, cauchy convergence, boundedness of continuous functions, differentiability of series, with proofs, in high school AP? in college calc? in college honors calc?
Although I'm not studying mathematics, we have covered all of this in my first year of engineering, in Analysis I and Analysis II. We didn't really construct the real numbers out of the rationals but did see the fundamental properties such as completeness and proved some fundamental theorems based on this (such as the lub-property, Bolzano-Weierstrass, Heine-Borel). From what you mentioned earlier, we also saw differentiating/integrating power series term wise, proving Riemann-integrability for continuous function using uniform convergence, we saw the inverse function theorem (without proof) and proved the implicit function theorem. And then a lot more which hasn't been mentioned before of course.

The way I understand it from this topic, this seems to be quite unordinary at other places, for a first year university, non-math direction. I must add though that the engineering studies at university level here in Belgium are a lot more theoretical (and mathematically founded) compared to most other countries - which I'm glad of seeing that nearly went for mathematics :smile:
 

Similar threads

  • STEM Academic Advising
2
Replies
43
Views
4K
  • STEM Academic Advising
Replies
4
Views
1K
  • STEM Academic Advising
Replies
4
Views
2K
  • STEM Academic Advising
Replies
2
Views
2K
  • STEM Academic Advising
Replies
3
Views
1K
  • STEM Academic Advising
Replies
14
Views
1K
  • STEM Academic Advising
Replies
3
Views
896
  • STEM Academic Advising
Replies
3
Views
1K
  • STEM Academic Advising
Replies
9
Views
2K
  • STEM Academic Advising
Replies
11
Views
522
Back
Top