Full Idea
Fictionalism takes an epistemology of the concrete to be more promising than concrete-and-abstract, but fictionalism requires an epistemology of the actual and possible, secured without the benefits of model theory.
Gist of Idea
Fictionalism eschews the abstract, but it still needs the possible (without model theory)
Source
Stewart Shapiro (Philosophy of Mathematics [1997], 7.2)
Book Reference
Shapiro,Stewart: 'Philosophy of Mathematics:structure and ontology' [OUP 1997], p.223
A Reaction
The idea that possibilities (logical, natural and metaphysical) should be understood as features of the concrete world has always struck me as appealing, so I have (unlike Shapiro) no intuitive problems with this proposal.