AbstractThis essay will be concerned with an evaluation, modification, and critique of van Inwagen’s famous argument for the existence of fictional characters. In the first section a reconstruction of the original argument will be provided, and three different challenges for this version of the argument will be pointed out. The first challenge concerns van Inwagen’s commitment to first-order predicate logic as a canonical language for the formal representation of truth-conditions of assertoric sentences in natural languages, and the problematic semantic complexity of van Inwagen’s original example sentences of quantifications about fictional objects. The second challenge concerns his commitment to a Quinean conception of ordinary language quantifications that conceives of first-order quantifiers as existentially loaded. The third challenge concerns van Inwagen’s tendency to interpret our ordinary intuitions about the truth-values of specific sentences as intuitions about the truth-values of the literal contents of these sentences. In the second section, a more detailed investigation of these three challenges will be provided; and a modified, and less problematic, alternative version of the argument will be proposed. It will be shown that the truth-conditions of van Inwagen’s original example sentences cannot be adequately represented on the basis of first-order predicate logic. I will propose alternative and less complex example sentences that are sufficient for the required purpose. Additionally, a reformulation of the argument will be proposed that avoids a commitment to a specific sort of formal framework. After that, it will be shown why the assumption of a Quinean conception of quantification unnecessarily increases the burden of proof. A reformulation of the argument will be proposed that avoids the commitment to a Quinean conception of quantification. Furthermore, I will make a third and final adjustment of the argument that allows us to remain neutral concerning the specific status of our truth-value intuitions concerning the proposed example sentences of generalisations about fictional objects. In the third and final section, three possible responses of an irrealist concerning fictional objects will be evaluated. The first option makes use of a recent semantic analysis of the modifying adjective ›fictional‹ proposed by Sainsbury. According to this analysis, a sentence like ›There are fictional mice that talk‹ is semantically equivalent with the claim ›There are fictional works according to which it is the case that at least one mouse talks‹. The second option additionally makes use of Sainsbury’s conception of spotty scope for ordinary language quantifiers and other related sentential operators. It will be shown why both options cannot account for the desired true readings of our example sentences. Finally, a third solution will be proposed and defended. This solution makes use of a substitutional interpretation of specific fictional generalisations based on a negative free logic to undermine the modified argument. It will be shown how this solution allows us to provide a correct analysis of the desired readings of our example sentences of generalisations about fictional characters. After that, two problems of this account will be discussed. The first problem concerns the extension of the proposed strategy to more complicated and sophisticated example sentences. The second problem concerns the independent motivation of a substitutional treatment of at least certain natural language quantificational expressions.