Structural and functional analysis of four-word lexical bundles in public debate

The present study aimed to investigate and analyze the structures and discourse functions of four-word lexical bundles employed in public debate. To achieve these aims, 49 episodes from the Intelligence Squared U.S. Debates podcast from the period of 2014 – 2019 were collected to build the Pub-D corpus. The list of four-word lexical bundles was generated with the use of AntConc software. The structures and discourse functions of the bundles were analyzed and classified according to Biber et al.’s (2004) taxonomy. The results revealed that the most prevalent structure used in a public debate was lexical bundles incorporating verb phrase fragments. Regarding the discourse functions, most bundles were classified as stance bundles to convey the speaker’s point of view on their topic, intention, prediction, and desires. This might yield some pedagogical implications on how these debates appeal to the audience’s beliefs during a discussion of an important issue.