Higher logarithms (Q1813745): Difference between revisions

The classical dilogarithm \(L_ 2\) was shown to be an important tool in many problems of arithmetics, \(K\)-theory, group cohomology and so on. It is the second term of a row \(L_ 1\) (=usual logarithm), \(L_ 2,\ldots\) --- The authors point out that they ``propose a new approach to constructing higher logarithms \(L_ p\), \(p=0,1,2,\ldots\)'' which will produce what they believe ``should be the true generalizations of the logarithm and the dilogarithm.'' Their construction of \(L_ p\) ``appears as a component of an interesting cocycle whose class in the Deligne- Beilinson cohomology \(H_ D^{2p}(G^ p_ \bullet,\mathbb{Q}(p))\) of a certain simplicial space \(G^ p_ \bullet\) (of Zariski open subsets of various Grassmannian manifolds) is a kind of universal \(p\)-th Chern class.'' The content of the paper includes the following sections: multivalued differential forms, the Grassmannian complex, higher Albanese manifolds, rational \(K(\pi,1)\)-spaces and the existence of the 3- logarithm, symmetry of the Grassmannian, non-triviality and indecomposability of the 3-logarithm, real Albanese manifolds, generalized Bloch-Wigner functions. The main difficulty lies in an existence theorem for the functions \(L_ p\).