### Kosmann-Schwarzbach & Weinstein on Lie Algebroid Classes

#### Posted by Urs Schreiber

On October 24 Yvette Kosmann-Schwarzbach will give a talk in the Hamburg math colloquium on

Modular classes and relative modular classes in Lie algebroid theory

based on work with Alan Weinstein

Y. Kosmann-Schwarzbach & A. Weinstein
*Relative modular classes of Lie algebroids*

math.DG/0508515

and Camille Laurent-Gengoux

Y. Kosmann-Schwarzbach & C. Laurent-Gengoux
*The modular class of a twisted Poisson structure*

math.SG/0505663 .

A Lie algebroid # is to a Lie groupoid as a Lie algebra is to a Lie group #.

On each Lie algebra $g$ we have a linear form coming from the trace of the adjoint action of $g$ on itself

This is a 1-cocycle in Lie algebra cohomology.

This construction may be generalized to Lie algebroids. The corresponding 1-class in Lie algebroid cohomology is called the **modular class** of the Lie algebroid.

Technically, the modular class of a Lie algebroid $E$ is a section $\xi$ of $E^*$ which satisfies, for all sections $x$ of $E$ the equation

where $\omega$ and $\lambda$ are nowhere-vanishing sections of $\wedge^\mathrm{top} E$ and $\wedge^\mathrm{top} T^* M$, respectively. (See p. 4 of the first paper cited above for details.)

There are “twisted” versions of Lie algebroids, where the familiar equations only hold up to higher coherence. These should be nothing but 2- (and higher) algebroids, but are not usually addressed this way.

The most prominent example is the **Courant algebroid** #. (I once thought I convinced myself that the Courant algebroid over a point is nothing but the Baez-Crans Lie 2-algebra $g_k$ #.)

There is something some people call a Dirac structure on a Courant algebroid. Kosmann-Schwarzbach et al. call this a **twisted Poisson structure**.

In the second of the above papers a notion of modular class for such a “Lie algebroid with twisted Poisson structure” is defined. One of the main points of these papers is to relate this definition to a generalization of the ordinary modular class of a Lie algebroid called the **relative modular class**.

Maybe I’ll discuss more of the details after I have heard the talk.