Résumé
Argumentation provides a formalism consisting of arguments and attacks/supports between these arguments and can be used to rank or deduce justified conclusions. In multi-agent settings, where several agents can advance arguments at the same time, understanding which agent has the most influence on a particular argument can improve an agent's decision about which argument to advance next. In this paper, we introduce an argumentation framework with authorship and define new semantics to account for the impact of the agents on the arguments. We propose a set of desirable principles that such a semantics should satisfy, instantiate such semantics from two popular graded based semantics, and study to which extent these principles are satisfied. These semantics will allow an observer to identify the most influential agents in a debate.