The main purpose of this essay is critical. I focus on Robin Jeshion’s (2002; 2004; 2010) theory of singular thought, and I offer three objections to her Significance Condition for the creation of mental files. First of all, this condition makes incorrect predictions concerning singular thoughts about insignificant objects. Second, it conflicts with a theoretical aim mental file theories usually have, that of accounting for our ability to track discourse referents. And third, it appeals to a vague notion where a clear-cut notion is needed. In the final section, I suggest that there are more plausible alternatives to the Significance Condition that the mental file theorist could appeal to, and which do not face the problems mentioned.