There has been increasing interest in utilizing corpora in linguistic research and pedagogy in recent years. Rhetorical organization of different sections of research articles may appear similar in various disciplines, but close examination may show subtle differences nonetheless. One of the features that has been at the center of attention especially in recent years is the idiomaticity of a discourse which can be examined from the perspective of multi-word units captured by the automatic retrieval of lexical bundles. This study takes a corpus-based approach for the identification of lexical bundles. A corpus of 801,894 words from 790 articles was collected. In order to fulfill the purposes of the present study, ABBYY FineReader 10 professional edition, Total Assistant, Antconc 3.2.3, and WordSmith Tools 5 were used to identify lexical bundles. Then these bundles were classified structurally and functionally based on the presented taxonomies in the literature. The results of the current study indicated that the writers of medical research articles mostly rely on text-oriented bundles in the discussion section of research articles to establish academic discourse.
