ABSTRACT
Prefabricated structures referred to as poly-word strings or lexical bundles are recognized as the fundamental building blocks in academic discourse. They build fluency in academic discourse, helping to shape coherence and meaning in a speech or text. For decades, lexical bundles have captivated a considerable amount of attention in corpus-based research in English for Academic Purposes. However, in the Pakistani context, less research has been carried out to explore lexical bundles in academic discourse. Therefore, the present study aimed to examine prefabricated structures employed in the scientific research abstracts of theses and articles from social and physical sciences. A specialized corpus was built from the 1200 research abstracts from social and physical sciences. AntConc software was utilized for the extraction of lexical bundles. Biber structural model of lexical bundles was employed as a theoretical framework. The study found 42 four-word lexical bundle types and 659 tokens in soft sciences. On the other side, the study found 55 four-word lexical bundle types and 950 tokens in physical sciences. Besides, this corpus-based study found 11 common core four-word lexical bundles in social and physical sciences. Moreover, the present study also contributed some new structures to the respective structural model of lexical bundles. If Pakistani science scholars are introduced with these lists of lexical bundles obtained from this study, their rhetorical practice of writing research abstracts can be significantly improved. The results yielded from this study can be substantially used for ESP and EAP further to enhance the academic discourse of the research scholars.
