Nations united through discourse: a corpus analysis of UN general assembly addresses

The overarching goal of this dissertation is to investigate the linguistic features of United Nations (UN) General Assembly (GA) General Debate addresses delivered in English during 2015. In order to describe UNGA addresses with the greatest accuracy, breadth, and depth, the study explores typical linguistic features across texts as well as variation within the register. Thus, the dissertation also examines the geographic, political, economic, and social profile of the countries represented in the UNGA in order to identify any country-related characteristics associated with systematic linguistic variation. The dissertation uses two corpus-informed methodologies: register analysis and Corpus-Assisted Discourse Studies (CADS), with a secondary goal of discerning the ways in which methodology affects the nature of findings. The study is based on a corpus of all 92 UNGA addresses delivered in English during the General Debate in 2015. In the examination of linguistic features across UNGA addresses, the UN texts are contrasted with the British National Corpus as well as four comparable registers: prepared speeches, spontaneous speeches, official documents, and conversation. The analysis begins with a description of situational characteristics for the UNGA and the four comparison registers. Then keywords and Multi-Dimensional scores for the five principal dimensions found in Biber (1988) reveal the lexical and grammatical features that characterize the UNGA. The second part of the investigation looks at variation among UNGA addresses, with texts divided into groups based on country-specific characteristics such as region and Gross Domestic Product. Keyword analysis, key feature analysis, and Multi-Dimensional analysis are used to identify systematic variation of linguistic features based on country group. The final analysis is a CADS study of three frequent and politically salient terms (terrorism, Security Council, and women) with keyword and collocational analyses to determine how the terms are conceptualized generally in the UNGA and also whether any variation can be identified for country groups. Results show a remarkable consistency across the UNGA not only in what lexical and grammatical features are preferred and dispreferred but also in how they are used. Lexical features reveal the importance of topic, purpose, and discourse structure. Grammatical features are greatly influenced by written-to-be-spoken production circumstances and purpose. In spite of the large number of features (126) and country groups (31) explored, very few exhibit any systematic patterns of variation. The variation that does occur is primarily associated with level of development and region. The CADS analysis supports the register analysis, showing that UNGA texts tend to conceptualize politically charged terms in similar ways with only occasional variation based on country group. These findings have important implications for the field of political discourse analysis, furthering our understanding of the linguistic features of international organization spoken discourse. In addition, the CADS analysis demonstrates some important methodological considerations, with consequences for the design of future corpus-informed research