Collections of data can be visualized as discrete portions, subsets, or branches of the Vocabulary tree. A parent entity is associated with a set of child entities, which are called call elements of the collection. The collection of pilots can be illustrated as:

Figure 1. Visualization of a collection of pilots

In this figure, the aircraft entity is the parent of the collection, while each pilot is a child element of the collection. As you saw in the role example, this collection is expressed as aircraft.pilot in the Corticon Rule Language. It is important to reiterate that this collection contains scope. You are seeing the collection of pilots as they relate to this aircraft. Or, more simply, you are seeing a plane and its 2 pilots, arranged in a way that is consistent with the Vocabulary. Whenever a rule exists that contains or uses this same scope, it also automatically evaluates this collection of data. And, if there are multiple collections with the same scope (for example, several aircraft, each with its own collection of pilots), then the rule automatically evaluates all those collections as well. In the Corticon lexicon, evaluate has a different meaning than fire. Evaluate means that a rule's scope and conditions will be compared to the data to see if they are satisfied. If they are satisfied, then the rule fires, and its actions are executed.

Collections can be much more complex than this simple pilot example. For instance, a collection can include more than one type or level of association:

Figure 2. Three-level collection

This collection is expressed as parent.child.grandchild in the Corticon Rule Language. 

Note: The parent and child nomenclature is a bit arbitrary. Assuming bidirectional associations, a child from one perspective could also be a parent in another.