The last decade has seen the advent of the agent paradigm as a reference model in several fields of research, mainly but not limited to artificial intelligence and distributed systems. In open and distributed environments, where most facts are not known at all, the agent metaphor proves particularly useful if agents are able to autonomously perform some form of reasoning, possibly obviating knowledge incompleteness by means of hypotheses assumed on the unknown facts. A suitable mechanism to deal with incomplete and multiple knowledge is abductive reasoning. The aim of this paper is to describe LAILA, a language that can be used by logic-based agents capable of abductive reasoning, by enabling them to express at a high level several ways to join and coordinate with one another. In particular, we considered collaboration and competition as possible interaction patterns in the abductive reasoning that must be carried out by multiple agents. Syntax and operational semantics of the LAILA language are given along with a clarifying example; a section is also devoted to a brief description of the current LAILA implementation.