Where to begin... :) This gets to the internal concept of 2-phase loading of entity data and how specifically it is done. Essentially for composite values, the process of hydration results in a tuple array comprising the composition "atomic" values. For example, a Name component's hydrated state might look like ["Steve", "L", "Ebersole"]. There are times when we need to be able to extract individual pieces out of the hydrated tuple array. For example, for an entity with a composite identifier part of which is an association (a key-many-to-one) we often need to attempt 2-phase processing on the association portion of the identifier's hydrated tuple array. This contract allows us access to portions of the hydrated tuple state.