Allow subvocabularies in mx

[Broes De Cat]

To allow the theory / term / structure to have subvocabularies of each other, one solutions is to generalize what can be looked-up in a structure.

Currently, it is possible to have a predicate which does not occur in a structure itself, although it can be mapped to a predicate which is interpreted in the structure. This is a quite complex solution which might make the overall system too complex. It might be a better choice to up-front change the vocabularies (internally) and then continue on with identical vocs. It depends a bit on what our view on symbol sharing is.

[Broes De Cat]

Proposal for solution (complex approach) can be found in branch allowsubvoc_in_mx