MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

We learn in complex analysis class how to find the winding number of the contour $\Gamma$ around the origin. \[ n = \frac{1}{2\pi i} \oint \frac{dz}{z} = \frac{1}{2\pi i} \oint d(\log z) = \frac{1}{2\pi } \oint d\theta \] My goal had been to count the number of self-intersections of curves. I guessed some integral over a torus $\Gamma \times \Gamma$ which would have a pole-like object whenever $z_1 = z_2$. In the back of my mind, I worried maybe integrating along $t_1 = t_2$ would have zero contribution. \[ \frac{1}{2\pi i} \oint \oint \frac{dz_1 dz_2}{z_1 - z_2} \]

Inspired by Kontsevich's integral for knots (and some more recent papers), I learned of something that comes close \[ \frac{1}{2\pi i} \oint \oint \frac{dz_1 - dz_2}{z_1 - z_2} = \frac{1}{2\pi i} \oint \oint d \log(z_1 - z_2) = \frac{1}{2\pi i} \oint \oint \arg (z_1 - z_2) \] So formulas like these are measuring how the chords of curves wind around each other. This seems to be known as the Whitney invariant for plane curves, counting signed self-intersections.

alt text

Is there a way to get self-intersections all of the same sign? This must be related the Vassiliev invariants as well, but I'd like to focus on plane curves.

share|cite|improve this question
This measure blows up when the curve is self-tangent, since if you go one way it has 2 intersection and if you go the other way it has 0. Therefore, you need to integrate something that has a pole or is otherwise not well-defined when the curve is self-tangent. This seems hard, because for it to be self-tangent $(z_1-z_2)$ and $d(z_1-z_2)$ both have to be zero. – Will Sawin Jun 1 '12 at 17:02

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.