This hierarchy does not contain any conflicts where one agent complains about not having enough knowledge to be able to do its job.
Such situations are possible between estimator-selector pairs, and evaluator-selector pairs. Suppose a selector chooses a material. Later an estimator, that is supposed to produce an estimate for another parameter, needs to use the value of the material parameter, but does not have knowledge about the particular material chosen.
Alternatively, an evaluator that is supposed to evaluate the particular material does not know about that material and therefore cannot evaluate it.
These situations are neither criticism conflicts, nor incompatibility conflicts. When an estimator or an evaluator does not know about the value produced by an estimator, learning should take place. The selector should be notified of the situation by a complaint message from the agent with the incomplete knowledge. Then, the selector should send information about its choice to the complaining agent. This is only one of the many possibilities for learning in the SiFA paradigm. Learning in multi-agent systems is a future research direction discussed in section 10.4.