MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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

The number $12$ (or, probably we shall say Bernoulli numbers in general) appears in many places in Mathematics, sometimes leading to unexpected connections between different topics.

For instance, some time ago there was a very interesting explanation for

1) its occurrence in the Todd class


2) its occurrence in the Euler-Maclaurin formula

in terms of Riemann-Roch for toric varieties, as explained in:

Euler-Maclaurin formula and Riemann-Roch


My question is, will there be some relation between 1) and

3) its occurrence in the Baker-Campbell-Hausdorff formula.

I guess this might be related to some explicit local expressions in some method of proving the index theorem on Lie groups, or even the Duflo map (which I don't really understand).

Thank you very much.

share|cite|improve this question
Isn't this a numerology question?! At least you don't ask why 12 is close to an integer... :-) – Wadim Zudilin Jul 15 '10 at 7:27
Hi Wadim, I edited the question make it somehow more "feasible" :-) – Bo Peng Jul 15 '10 at 7:42
A baker's dozen is 13. I don't know about Campbell and Hausdorff. – Tom Goodwillie Jul 15 '10 at 11:11
Possibly related… – SandeepJ Jul 15 '10 at 14:11
Have you looked at the Wikipedia article, Please, clarify in what way the explanation there involving the expansion of $\operatorname{ad}(X)/(e^{\operatorname{ad}(X)}-1)$ found there is lacking? – Victor Protsak Jul 15 '10 at 17:56

The answer to your question is the following: given two non-commutative variables $x$ and $y$ one has $$ log(e^xe^y)=x+e^{ad_x}\frac{ad_x}{e^{ad_x}-1}(y)+O(y^2) $$

It is not the appearance of $12$ that is intriguing, but the appearance of the Todd series in algebraic geometry. It suggests that there is a group hidden somewhere... and this is indeed the case. This group is the derived loop space of your favorite algebraic variety $X$, and its tangent Lie algebra is the shifted tangent sheaf $T_X[-1]$, with Lie bracket given by the Atiyah class (the fact that the Atiyah class gives rize to a Lie structure was discovered by Kapranov).

The universal enveloping algebra of this Lie algebra is the Hochschild complex of $X$. One then gets a nice dictionnary between the Lie side and the algebraic geometry side. E.g.:

  • any object in the derived category of $X$ turns out to be a representation of this Lie algebra.

  • Poincare-Birkhoff-Witt is Hochschild-Kostant-Rosenberg.

  • the Duflo isomorphism is the Kontsevich-Caldararu isomorphism between the Harmonic and Hochschild structures.

  • there is also an relation between closed embeddings in algebraic geometry and inclusions of Lie algebras.

  • ...

share|cite|improve this answer
In item 1 "derived category of X " do you mean current sheaves on X? PS very useful answer! – Alexander Chervov Dec 5 '12 at 19:14
Coherent sheaves – Alexander Chervov Dec 5 '12 at 19:43
I actualy mean quasi-coherent sheaves (the action of $T_X[-1]$ on a given one $E$ is given by the Atiyah class of $E$). – DamienC Dec 5 '12 at 20:07

Your Answer


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

Not the answer you're looking for? Browse other questions tagged or ask your own question.