Using text-mining technology, this article traces diachronic changes of narrativity in the genre of the British and American novel from the 18th-century to the mid-twentieth-century. Recently introduced to the various fields of the humanities, text-mining is a viable way in which quantitative analyses can be used in literary studies. To open wide the possibility of quantitative anlayses of the genre of novel, this article shows how text-mining is used and innovates the study of novels through sample analyses of Henry Fielding, Daniel Defoe, James Joyce, and Ernest Hemingway. The results show the distinctive differences in narratorial control (first person pronoun), syntactic simplicity, and other grammatical and stylistic features. For instance, in the samples, Hemingway enhanced narrativity by increasing word concreteness and James Joyce by syntactic simplicity, whereas Defoe enhanced it by increasing referential cohesion and Fielding by deep cohesion.