Some conjectures and results about multizeta values for \(\mathbb F_q[t]\) (Q963002): Difference between revisions

Let \(a,b>1\) be integers. Then the product \(\zeta(a)\zeta(b)\) can be expressed as a sum of multi-zeta values using the sum shuffle identity \(\zeta(a)\zeta(b)=\zeta(a+b) + \zeta(a,b) + \zeta(b,a)\). \textit{D. S. Thakur} [Function field arithmetic. River Edge, NJ: World Scientific (2004; Zbl 1061.11001)] has introduced and studied characteristic \(p\) function field analogues of multi-zeta values. As in the classical case, one expects that there are identities expressing the product of two Goss zeta values a sum of Thakur multi-zeta values. The naive analogue of the sum shuffle identity however fails, and the correct identities are expected to be much more involved. In the present paper, conjectures are stated about the precise form of these identities and many computer verifications are presented. Recently \textit{D. S. Thakur} [Int. Math. Res. Not. 2010, No. 11, 1973--1980 (2010; Zbl 1198.11077)] has used different methods to show that the product of two zeta values is a sum of multi-zeta values.