Math H110A, Fall 1999, TTh 11-12:30

Professor Allen Knutson
Text: Linear Algebra, An Introductory Approach, Charles W. Curtis, Springer UTM
Course web page: http://www.math.berkeley.edu/~allenk/courses/fall99/h110a.html
Office hours: 10-11 T/Th, 11-12 W

Welcome to Math H110A, Honors Linear Algebra! I am designing this course for a look beyond the formal manipulations with matrices to real understanding of the geometry and composition of vector spaces and linear transformations. On this page:

  • General remarks about this course
  • Things we will definitely do, will probably do, and might do this term
  • How the course will be run
  • Math class is hard!
  • On the first day of class I handed out this survey to get a sense of what the people entering this class already know.

    General remarks about this course

    "Choosing coordinates on a vector space...is an act of violence"
    -- Hermann Weyl

    A central goal, permeating this course, will be to develop a sense of which linear algebra questions should be approached from a highbrow, coordinate-free manner (almost all), and which should be done down-and-dirty with matrices (almost none). Is this just mathematical snobbery? No (at least, not just that); the highbrow proofs, once understood, are simpler and less prone to error, and always much much shorter.

    Things we will definitely do, will probably do, and might do this term

    In past incarnations (by other professors) this course has been a mix of some absolutely obligatory results - I'm thinking Jordan Canonical Form here - and special topics, such as vector spaces adorned with a finite group of symmetries, or set-theoretic questions about infinite-dimensional vector spaces. Here is a list of what I consider essential, what special topics I fully intend to explore, and more fanciful topics we may attack if time permits. All page numbers refer to the sixth printing of Curtis' book, yellow with the ellipse on the cover.

    Essential topics, largely in the order we will do them

  • *Definitional stuff
  • Vector spaces (other than R to the n)
  • Subspaces
  • Span, linear independence, bases, dimension
  • Vector spaces over the field {0,1} with 1+1=0
  • *Linear transformations, and their representation as matrices
  • *Traces and determinants
  • The interpretation of determinants as (signed) volumes
  • The geometry of 2x2 matrices
  • The Jordan canonical form of a square matrix
  • Dual spaces and tensor products
  • The KAN, KP, and BwB decompositions of invertible matrices
  • Projective spaces
  • "*" means I hope you've seen this once already.

    Less essential topics, still important for much subsequent mathematics, that I fully expect us to get to

  • Inner products: orthogonal, unitary, and Hermitian matrices
  • The principal eigenvalues of symmetric and Hermitian matrices
  • Kleinberg's algorithm for web searches
  • Graph embedding
  • Minimax inequalities
  • Uniqueness properties of the Jordan canonical form
  • Some other cool advanced linear algebra things, some of which will fit in somewhere

  • Normal modes of finite systems of springs
  • Modules over rings that aren't fields
  • Classification of finite abelian groups
  • Modules over the matrix algebra M_n(D)
  • The Artin-Wedderburn theorem and representations of finite groups
  • Tensor algebras, symmetric algebras, exterior algebras, Clifford algebras
  • Division algebras over the reals: R,C,H
  • The classical groups
  • Matroids
  • Applications of projective geometry to computer graphics
  • Honeycombs
  • How the course will be run

    Like so many before it, this class will have homework, a midterm, and a final. Homework will be assigned on Thursdays, due the following Thursday, and collected in class. Having myself put off many a homework to the last minute and handed it in at the end of a totally-ignored class, I will collect homework at the beginning of Thursday's class. The best time to ask me questions about it is face-to-face in the couple of minutes before Tuesday's class, such that I can make an announcement if there's some egregious error or something.

    I would rather assign interesting problems, whose results we will make use of and generalize in the classes to come. In particular, many problems will probably become very easy shortly after they are due. So late homework - later than the beginning of class - is only going to count 50%. If you know in advance you won't be able to make it to Thursday's class you can slip your homework under my office door.

    The homework will be 50%, the midterm 20%, and the final 30% of your grade. I have no idea what percentage corresponds to what letter, but will announce estimations after the midterm.

    Math class is hard!

    -- a swiftly discontinued Talking Barbie

    This being an honors class, I know you're here to really learn some mathematics, and I will do my best to make sure you learn as much as is practicable by the end of the course. This is of course subtler than throwing the maximum amount of material at you -- the hard part (for me too) is making sure that at the end of the course you remember and understand everything that has come before.

    One thing it does mean is that, if I'm doing my job right, the homeworks should challenge everyone. Too many people getting perfect grades on too many homeworks means that the homeworks are lame. For those of you who haven't taken classes tuned to this sort of standard, this will be disconcerting for a while, and all I can do is assure you that I will have it in mind when converting percentages to letter grades. Those of you who have know how much more can be gleaned from such a course.