We consider the possibility to explain manifestations of collective intelligence via the presence of structure in the set of tasks may being fulfilled. A distributed system corresponds to a lattice of tasks that the system can fulfill. On this lattice, we define an external multiplication operation for lattice elements and respective operations of linear logic. The implication operation is used to choose the version of system reconfiguration, i.e., choose an executor for a new task. We show that this choice is to a large extent defined by the structure of the lattice of tasks and the definition of the multiplication operation on it. The proposed reconfiguration method is applied to an example of operation for a group of janitor robots.