J.M. Isidro
Abstract: Superimposed D-branes have matrix-valued functions as their transverse coordinates, since the latter take values in the Lie algebra of the gauge group inside the stack of coincident branes. This leads to considering a classical dynamics where the multiplication law for coordinates and/or momenta, being given by matrix multiplication, is nonabelian. Quantisation further introduces noncommutativity as a deformation in powers of Planck's constant. Given an arbitrary simple Lie algebra and an arbitrary Poisson manifold, both finite-dimensional, we define a corresponding C*-algebra that can be regarded as a nonabelian Poisson manifold. The latter provides a natural framework for a matrix-valued classical dynamics.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the existence conditions of a specific type of "sum of squares" formula. Specifically, the authors explored whether there exists an identity of the form:
\[
(x_1^2 + \cdots + x_r^2)(y_1^2 + \cdots + y_s^2) = z_1^2 + \cdots + z_n^2
\]
on an arbitrary field \( F \), where \( z_i \) are bilinear expressions of \( x \) and \( y \). Such an identity is called a "sum of squares" formula of type \([r, s, n]\). The classical theorem uses topological K - theory to prove that if such a formula exists on the real number field \( \mathbb{R} \), then certain powers of \( 2 \) must divide certain binomial coefficients. Although it is known that this result holds for all fields of characteristic \( 0 \), the case of characteristic \( p \) has been unresolved. In this paper, the authors use algebraic K - theory instead of topological K - theory to prove that this result holds for all fields.
The main contribution of the paper lies in extending this result to fields of non - zero characteristic, thus filling the gap in previous research. Specifically, the main theorem of the paper shows that if a "sum of squares" formula of type \([r, s, n]\) exists on a field \( F \) whose characteristic is not \( 2 \), then for \( n - r < i \leq \left\lfloor \frac{s - 1}{2} \right\rfloor \), \( 2^{\left\lfloor \frac{s - 1}{2} \right\rfloor - i + 1} \) must divide \( \binom{n}{i} \). This conclusion specifically points out that a formula of type \([13, 13, 16]\) does not exist on any field whose characteristic is not \( 2 \), which was previously only confirmed on fields of characteristic \( 0 \).