The desirability of a corpus of online book responses

摘要

This position paper argues the need for a comprehensive corpus of online book responses. Responses to books (in traditional reviews, book blogs, on booksellers' sites, etc.) are important for understanding how readers understand literature and how literary works become popular. A sufficiently large, varied and representative corpus of online responses to books will facilitate research into these processes. This corpus should include context information about the responses and should remain open to additional material. Based on a pilot study for the creation of a corpus of Dutch online book response, the paper shows how linguistic tools can find differences in word usage between responses from various sites. They can also reveal response type by clustering responses based on usage of either words or their POS-tags, and can show the sentiments expressed in the responses. LSA-based similarity between book fragments and response may be able to reveal the book fragments that most affected readers. The paper argues that a corpus of book responses can be an important instrument for research into reading behavior, reader response, book reviewing and literary appreciation.