# Poisson bracket of coordinates

I just derived that in local coordinates (it suffices to centre) around $0$, that

$$\{f,g\}(x)=\sum_{i,j}\{x^i,x^j\}\frac{\partial f}{\partial x^i}\frac{\partial g}{\partial x^j}$$

only using the axiomatic properties of the Poisson bracket $$\{*,*\}:C^\infty(M)\times C^\infty(M)\to C^\infty(M)$$ 1) The conjugate of Poisson is the Poisson of conjugates, 2) bracket is bilinear, 3)antisymmetric, 4) Leibniz rule is satisfied, and 5) the Jacobi identity is satisfied.

Essentially I took the Taylor series around $0$ then the first order terms clearly die out ($\{f,1h\}=\{f,1\}h+\{f,h\}$, so any constant is destroyed by the bracket), and the higher order terms die out because you get

$$\sum_{ij} \frac{\partial^2 f}{\partial x^i \partial x^j}\frac{\partial^2 g}{\partial x^i \partial x^j}\{x_i x_j,x_i x_j\}$$ for the Poisson bracket of the second order terms, third order terms have $\{x^i x^j x^k,x^i x^j x^k\}$ and so on, and obviously $\{g,g\}=-\{g,g\}=0$.

My question is what is (geometrically or otherwise) the meaning of $\{x^i,x^j\}$? Again, $\{x^i,x^i\}=0$ but what should that $mean$.

-
For example, I would like the answer to be meaningful enough to where if someone wanted to program this into a computer they could. (i.e. if someone had to model quantization deformation of a Poisson manifold). I don't have to do this, but I believe that if something can be explained in such a lucid way as to be programmable then it's sufficiently clear! –  Squirtle Apr 26 '13 at 2:04

The equation you got for the Poisson bracket shows that the functions $\{x^i,x^j\}$ piece together to give the coefficients for a global anti-symmetric tensor field (called a bivector field) $$\Pi = \{x^i,x^j\} \frac{\partial}{\partial x^i} \wedge \frac{\partial}{\partial x^j} \in \Gamma(M, \Lambda^2 TM).$$ We recover the Poisson bracket by the equation $$\{f,g\} = \Pi(df,dg).$$
If the field $\Pi$ is non-degenerate (i.e. non-degenerate as a bilinear form $T_x^* M \times T_x^*M\to \mathbb R$) then we have an isomorphism $T^* M \to TM$ and we can convert $\Pi$ to a differential 2-form $\omega$. Then $\omega$ is a symplectic structure (closedness of $\omega$ is equivalent to the Jacobi identity for $\{\cdot,\cdot\}$).
I'm certain this answer is completely correct.... But I don't find it useful, in the sense that I'm very green to this topic, so the other stuff you wrote doesn't really explain what $\{x^i,x^j\}$ means. Is there a geometric meaning to it? –  Squirtle May 7 '13 at 22:09