"Know" and "think" occur in overlapping linguistic contexts but differ in factivity: only in using "know" does the speaker presuppose the truth of the complement. In contrast to the relative wealth of knowledge about when children acquire factivity, not much is known about how the child comes to this understanding. Using corpus methods, we examine children's input to determine which distributional cues (e.g., pragmatic, syntactic) might be informative about the (non-)factivity of "know" and "think". Our results indicate that certain cues could be useful in differentiating the verbs (i.e., types of complements that are embedded) while others could be useful in cuing factivity (i.e., the types of discourse moves that they are used to make). These findings will direct future behavioral experiments to determine the relationship between input and acquiring factivity.