ellipsis — the principled omission of linguistically recoverable elements — has been extensively documented as a key cohesive mechanism in English discourse. Yet the ways in which its frequency, type, and function differ across spoken and written modes remain incompletely understood, particularly in relation to the interplay between register, communicative purpose, and grammatical licensing. This article presents a comparative corpus-based study of ellipsis across four registers: spontaneous spoken conversation, scripted broadcast speech, literary prose fiction, and published academic writing. Drawing on a 500,000-word annotated corpus, the study analyses the distribution of VP ellipsis, clausal ellipsis, nominal ellipsis, gapping, sluicing, and stripping across these modes, and investigates the discourse functions each type serves within its host register. The results reveal that ellipsis is substantially more frequent and structurally diverse in spoken registers, where it fulfils interactional and processing functions, while written registers deploy it selectively for stylistic economy and information focus. These findings carry significant implications for discourse theory, corpus linguistics, and the teaching of both spoken and written English.
Khamza Panjiyevich Avazov (Wed,) studied this question.