One way to organize the information in a knowledge domain is with a tree structure. A tree is useful since it allows us to access the information in a manner that makes decision making faster, easier and more reliable. This structure is based on the idea that as humans we can tell the difference between things. For example we can tell the difference between dogs and cats; we can identify what characteristics they have in common with others in their natural group (both normally have four legs), and also how they are different from each other (cats have retractable claws, dogs bark). The tree structure gives us a useful graphic description of this parent-child hierarchy. This arrangement is also convenient for a computer which is able to track such relationships easily once instructed.
A simple tree with sample nodes
It also gives us a way to explore the nature of the nodes that make up the tree and the different traits that define them. We claim that each node consists of sets of traits in prototype and instantiated forms. The node 'Things' (those items in the real world with physical, tangible existence) may act as parent to two child nodes: living things and non-living things. In the first we find trees and animals, and in the second a wooden spoon or copper metal, for example. In the image such a tree is shown with nodes connected by dotted lines indicating that finer grained nodes may be inserted as required.
A simple node with traits
Considering the 'non-living things' node for the moment, let's say that in its role as parent it contains three traits among others: not alive, made of wood and made of metal. The characteristic of physical existence, tangibility and so on has been inherited from its parent 'Things'; also its own first trait has been extended (etymological meaning of trait is something that can be pulled out), however the other two are present but not yet extended. They become extended in the child nodes: copper being metallic and the spoon being wooden. The children inherit from the grandparent the trait of physical existence and not living from the parent, so there is no need to repeat these attributes in the child node. We arrange our parent nodes and define them in such a way that all nodes are connected to a parent, with the exception of the root node which is special.
We have claimed that each node contains a combination of extended and non-extended traits. The non-extended traits are in prototype form, meta-traits that will be used to define the children. Those already extended define the node itself, and were present in its parent in prototype form. In the leaves of the tree we find individuals; these nodes contain only extended traits. They have no children so they offer no prototype traits. They only have instantiated traits, in such a combination as to as to demonstrate their individuality.
In a similar way we can trace nodes back to the root of the tree, and ask ourselves what the root must look like. It should be a single node at its own level otherwise it would have siblings and a parent, in which case that parent becomes the root. The root has to start somewhere; in many tree structures it is just a hook to hang the whole tree on and we have no expectation of further functionality. But it can be co-opted into something more.
Following the pattern that each parent may contain prototype and extended traits, we see that, unlike intermediate nodes, since the root node has no siblings it has no need for any extended traits to set itself off against others at the same level. It can therefore be seen as a node consisting of only prototype traits. It is nothing, but has the potential to be everything. This is how some religions define God.