I discuss the question of ontological commitment of plural quantification. First, I introduce the language of Quantified Name Logic (QNL^e) and provide it with a set-theoretic semantics . Next, I compare QNL^e with Boolos's logic of plurals and prove that the expressive power of both languages (modulo set-theoretic semantics) is the same. Having done that, I discuss the question whether plural quantification commits its users to sets. I start with a "phony" argument to the effect that it does. It ultimately rests on the assumption that the logic obtained with the substitutional semantics is devoid of certain theoretically preferable features that QNL^e with set-theoretic semantics has. This is the assumption that I challenge. In order to do that I develop a substitutional semantics which avoids the objections raised before. First, I discuss two earlier attempts of providing plural quantification with ontologically innocent semantics. I present combinatorial semantics (introduced by Peter Simons) and indicate its weak points. Next, Rickey's natural models are defined. Unfortunately it is quite probable that Rickey's semantics does not avoid the objections against substitutional reading of quantifiers raised before. I develop a Kripke semantics for the logic of plurals (some of basic ideas stem from the considerations of Simons and Rickey). It is a modal interpretation, where the plural quantifier `exists_aphi' (suppose phi does not contain free variables other than a) is intuitively read as "it is possible to introduce a name a which would make phi substitutionally true". There are good reasons to claim that QNL^e with Kripke semantics has the same expressive power as QNL^e with set-theoretic semantics (as it turns out, for any set-theoretic model there is a Kripke model which agrees with it on all QNL^e formulas and the other way round: for any Kripke model there is a set-theoretic model which agrees with it on all formulas). Next, I explore the modal factor involved in this semantics by developing a two-sorted first-order modal logic of naming (MLN). I show that QNL^e with Kripke semantics can be embedded in the language of MLN. Since this embedding preserves models (i.e. models for QNL^e and MLN are the same, what changes is just the interpretation of symbols), it seems that the ontological commitment of QNL^e with Kripke semantics does not go beyond the ontological commitment of first-order (two-sorted) modal logic (with one relation constant). Since it is much less plausible that first-order modal logic commits one to abstract objects than that plural quantification does, this strengthen my case for the ontological innocence of QNL^e. The last section contains philosophical comments regarding the content of previous sections. I briefly compare my approach to Chihara's theory of constructibility quantifiers, discuss the issues pertaining to the ontological commitment, show how this approach avoids the usual objections against substitutional interpretation of plural quantifiers, and defend the view that there is nothing philosophically suspicious about restricting admissible models to complete naming structures (a notion to be defined).