Are Seven Words All We Need? Recognizing Genre at the Sub-Sentential Level
Philip M. McCarthy, Danielle S. McNamara · eScholarship (California Digital Library) · 2007
Genre recognition is a critical facet of text comprehension.In this study, we assess the minimum number of words in a sentence necessary for genre recognition to occur.Using corpora of Narrative, History, and Science sentences, we found that three experts in discourse psychology (demonstrating high agreement) accurately recognized the genre of over 80% of the sentences.This recognition generally occurred within the first seven words, with the highest accuracy for the Narrative genre.Thus, even very short and incomplete text can potentially activate text-structure knowledge and facilitate comprehension.In addition, we show that Narrative-like sentences are the most pervasive sentence type, with expert raters assigning 51% of misclassified sentences to the Narrative genre (again with high agreement between raters).In contrast, only 11% of misclassified sentences were assigned to Science.This study allows us to establish baseline expectations for skilled readers so that we can further examine differences in speed and accuracy of genre recognition as a function of reading skill.