Event and Agent Nominalisations Across Academic, Spoken, and Fiction Registers in Contemporary English: A Pilot Corpus Study of Clausal Density

Abstract

Nominalisation is a central device for packaging clausal information into noun phrases, thereby increasing clausal density and enabling the abstraction typical of academic prose. This article reports a pilot corpus study of nominalisation types across four English register samples drawn from the Corpus of Contemporary American English (COCA) and the British National Corpus (BNC): Academic, Spoken, and Fiction. Using formal diagnostics from generative work on nominalisations (Picallo 1991; Baker & Vinokurova 2009) as an annotation guide, but adopting a corpus-register perspective for explanation (Biber et al. 1999; Nichols 1989), 500 tokens were manually classified as Event nominalisations, Agent nominalisations, or complex nominal constructions. The results show a robust register contrast within this stratified sample: Academic writing overwhelmingly favours eventive nominalisations (about 80%), while spoken and fiction samples favour agentive nominals and show more surface noun-phrase complexity. A chi-square test confirms that the association between register group (academic vs non-academic) and nominal type is large in the sample. The discussion interprets these patterns as register-conditioned preferences for conceptual reification and agent suppression, while also foregrounding the theoretical and methodological limits of using corpus distributions as evidence for formal grammatical architecture.