EN
The article contains a brief presentation of Quine's arguments in favor of metamathematical realism and raises the problem of the dependence of ontological commitments of a theory on 'ideological' assumptions concerning the range of logical concepts. The problem is discussed by reference to Henkin and Boolos quantifiers. If these quantifiers are chosen and introduced as semantically primitive terms, fully understandable without recourse to set theoretical semantics and paraphrases presented as logical sentences with function variables and set variables, the resulting theory makes different ontological commitments from those that are made by Quine. The difference is illustrated by discussing a theory proposed by Hellman