Full Idea
Just as one can extend the domain of discourse to include solutions to the equation 'x^2=-1' so one can extend the domain of discourse to include objects that satisfy the condition 'x is the sum of the G's' or 'x is a temporal part of the object b at t'.
Gist of Idea
Just as we introduced complex numbers, so we introduced sums and temporal parts
Source
Kit Fine (The Question of Ontology [2009], p.164)
Book Reference
'Metametaphysics', ed/tr. Chalmers/Manley/Wasserman [OUP 2009], p.164
A Reaction
This thought lies behind Fine's 'Proceduralism'. I take it that our collection of abstracta consists entirely of items we have either deliberately or unthinkingly 'introduced' into our discourse when they seemed useful. They then submit to certain laws.