Endpaper: How To Compute Determinants
Endpaper: How To Compute Determinants
9
Endpaper:
How to Compute Determinants
Prof. Dennis Gaitsgory†
Harvard University
Cambridge, MA 02138
[email protected]
During one of my years in graduate school in Israel, I was a teaching fellow for a class on linear
algebra. I found the job annoying for two reasons: On one hand, the students were primarily non-math
majors. But more importantly, my class started at eight in the morning, which did not rhyme well with
my lifestyle at the time. As a result, I could not bring myself to prepare my section in advance. Instead
I improvised each time....
One day I found myself explaining determinants. “You know, for a generic matrix a determinant is
never zero. Somebody, give me an example of a matrix!” The class produced no reply. They were no
less sleepy than I was. In fact, not only were they asleep but they were suspicious as well. They did not
want to risk giving a matrix which by misfortune would have a zero determinant, with the gloomy title
of “degenerate” attached to it.
So I proceeded: “OK, let’s take the first matrix that comes to mind.”
1 2 3
4 5 6
7 8 9
I set about computing the determinant by the usual formula. I was never good with computations and,
once again, I was especially sleepy:
1 · 5 · 9 − 2 · 4 · 9 ± 3 · 4 · 8 + ....
It took me a good 10 minutes. And what a shock, the determinant was zero! “I must have made a
mistake,” I told the class. I ran through the calculations once more, checking every step. Another 10
minutes passed. Zero again!
I tried to save myself. “OK, but sometimes the determinant is zero. Sorry. But now let’s take a
really generic matrix.”
1 2 3 4
5 6 7 8
9 10 11 12
13 14 15 16
Another lengthy computation. . . .
At the end of that semester I was forced to enroll in a special seminar for delinquent instructors.
81