Full Idea
For Frege there is no need to specify the domain of the individual variables, which is taken as the totality of all objects. This contrasts with the standard notion of an interpretation, which demands that we first fix the domain.
Gist of Idea
Frege's domain for variables is all objects, but modern interpretations first fix the domain
Source
comment on Gottlob Frege (Begriffsschrift [1879]) by Michael Dummett - Frege Philosophy of Language (2nd ed) Ch.14
Book Reference
Dummett,Michael: 'Frege Philosophy of Language' [Duckworth 1981], p.475
A Reaction
What intrigues me is how domains of quantification shift according to context in ordinary usage, even in mid-sentence. I ought to go through every idea in this database, specifying its domain of quantification. Any volunteers?