Abstract
This research seeks to provide an insight into the identification and understanding of linguistic markers of Chief Executive Officer (CEO) hubris. It analyses spoken and written discourse samples of CEOs deemed to be hubristic and benchmarks the results against those of a sample of non-hubristic CEOs. In doing so it explores the hypothesis that the linguistic utterances of hubristic CEOs show consistent differences from the language produced by CEOs who have not been identified as possessing hubristic tendencies. This thesis presents a review of academic literatures pertaining to personality, hubris and natural language use. The review of these three domains leads to the conclusion that certain personality traits are antecedents to hubris and can be identified in one’s language use. The word count strategies are reviewed in depth as a framework for measuring hubris at-a-distance through the assessment of CEO’s linguistic utterances. In addition to word count strategies, this thesis also proposes a new approach – applying machine learning techniques to the analysis of language - for identifying CEO hubris. The research consists of a Pilot Study and three main studies (Study 1, Study 2, Study 3 (comprising two sub-studies, 3a and 3b)). It describes in detail the process, methods and materials used, summarises findings and explains the implications of the results obtained for further research into linguistic markers of hubris. For the purpose of this research, Hubris Syndrome is conceptualised as proposed by Owen and Davidson (2009), including all 14 proposed symptoms for Hubris Syndrome (Owen & Davidson, 2009). This research focuses explicitly on leaders who occupy or have occupied the position of CEO for a significant amount of time and were identified by other researchers, subject matter experts and media as having exhibited the features of Hubris Syndrome during their time in office. This research proposes several innovative techniques to identify the differences between hubristic and non-hubristic language, and documents subtle differences identified. Findings from this doctoral study suggest that the high use of impersonal pronouns, the total count of pronouns, auxiliary verbs, common verbs and tentative tone indicate CEO hubris. All in all, exploring if and how hubris’ symptoms manifests in CEO language use and what are characteristic features of hubristic discourse, contributes to wider research regarding the diagnosis and prevention of this phenomenon. This study seeks to mitigate the risk of potentially destructive CEO behaviour for the organisation and prevent organisational failures induced or aggravated by Hubris Syndrome.