In the language of science there are some theoretical (abstract) terms like molecule, momentum or space. Some of the empiricists are afraid of using them because they can implicate metaphysics (realism or nominalism). Rudolf Carnap in 'Empiricism, Semantics and Ontology' argues that beyond theoretical terms there is no metaphysics nor ontology because everything is reduced to the linguistic framework, in which he distinguishes internal and external questions we can ask. Internal questions and answers are restricted to the using of the new forms of expressions which are currently needed. External questions are often seen as questions of metaphysics but in fact, says Carnap, they are pragmatic questions concerning linguistic framework, such as: should we accept the term 'momentum' in the scheme of physical language? In this paper the author will analyse Carnap's arguments from the above-mentioned essay and consider whether he correctly solved the problem of the external question. Do really theoretical terms not implicate any ontology? To answer these questions he will confront Carnap's arguments with Willard Orman Van Quine's.