Abstract
ABSTRACT This paper introduces a method for computational analysis of move structures in abstracts of research articles and presents its pedagogical applications. In our approach, sentences in a given abstract are analyzed and labeled with a specific move in the light of various linguistic features. The method involves automatically gathering a large number of abstracts from the Web, manually assigning tags to sentences in a small set of the abstracts as well as to collocations extracted from abstracts, and automatically learning the relationships among words and move and the Markov model of move sequences. These phraseological and inter-move relationships are eventually used to label sentences of abstracts gathered from the Web. According to these labeled sentences, we present a prototype concordancer, CARE, which exploits the move-tagged abstracts in a digital environment. Independent evaluation on a set of real abstracts shows that the move-tagging method performs with high precision. Pedagogical trials show that the concordancer enriched with words, phrases and moves significantly enhance non-native speakers’ understanding of well-structured abstracts in writing. Our methodology clearly integrates computational, analytical framework and resources, leading to new ways of Web-based computer-assisted academic writing. Key words: Academic Writing, Abstract, Move Structure, Collocation, Concordancer