Natural language understanding depends heavily on assessing veridicality -whether events mentioned in a text are viewed as happening or not, but little consideration is given to this property in current relation and event extraction systems. Furthermore, the work that has been done has generally assumed that veridicality can be captured by lexical semantic properties whereas we show that context and world knowledge play a significant role in shaping veridicality. We extend the FactBank corpus, which contains semantically driven veridicality annotations, with pragmatically informed ones. Our annotations are more complex than the lexical assumption predicts but systematic enough to be included in computational work on textual understanding. They also indicate that veridicality judgments are not always categorical, and should therefore be modeled as distributions. We build a classifier to automatically assign event veridicality distributions based on our new annotations. The classifier relies not only on lexical features like hedges or negations, but also structural features and approximations of world knowledge, thereby providing a nuanced picture of the diverse factors that shape veridicality."All I know is what I read in the papers" Volume , Number speaker commitment, whereas qualified variants such as There are strong indicators that the cancer has spread or The cancer might have spread imbue the claim with uncertainty.We call this event veridicality, building on logical, linguistic, and computational insights about the relationship between language and reader commitment (Montague ). The central goal of this paper is to begin to identify the linguistic and contextual factors that shape readers' veridicality judgments. 1There is a long tradition of tracing veridicality to fixed properties of lexical items (Kiparsky and Kiparsky 1970;Karttunen 1973). On this view, a lexical item L is veridical if the meaning of L applied to argument p entails the truth of p. For example, since both true and false things can be believed, one should not infer directly from A believes that S that S is true, making believe non-veridical. Conversely, realize appears to be veridical, because realizing S entails the truth of S. The prototypical anti-veridical operator is negation, since not S entails the falsity of S, but anti-veridicality is a characteristic of a wide range of words and constructions (e.g., have yet to, fail, without). These basic veridicality judgments can be further subdivided using modal or probabilistic notions.For example, while may is non-veridical by the basic classifications, we might classify may S as possible with regard to S. 2 Lexical theories of this sort provide a basis for characterizing readers' veridicality judgments, but they do not tell the whole story, because they neglect the pragmatic 1 Our use of the term 'veridicality' most closely matches that of Giannakidou (1999), where it is defined so as to be (i) relativized to particular agents or perspectives, (ii) gradable, and (iii) general enough to co...