Full Idea
Russell proposed that descriptions be treated along with the quantifiers, which departs from Frege, who treated descriptions as proper names. ...the problem was that names invoke objects, and there is no object in failed descriptions.
Gist of Idea
Russell explained descriptions with quantifiers, where Frege treated them as names
Source
report of Bertrand Russell (On Denoting [1905]) by Gregory McCullogh - The Game of the Name 2.16
Book Reference
McCulloch,Gregory: 'The Game of the Name' [OUP 1989], p.42
A Reaction
Maybe we just allow intentional objects (such as unicorns) into our ontology? Producing a parsimonious ontology seems to be the main motivation of most philosophy of language. Or maybe names are just not committed to actual existence?