World Englishes, 2017 0883-2919 doi: 10.1111/weng.12257 Encapsulating Irish English in literature CAROLINA P. AMADOR-MORENO ∗ AND ANA MAR ´ IA TERRAZAS-CALERO ∗∗ ABSTRACT: This article examines the representation of Irish English in contemporary Irish writing, with particular reference to the narrative of the Ross O’Carroll-Kelly saga, a series of comic novels created by Irish writer Paul Howard which has enjoyed phenomenal success in Ireland, due in part to the author’s ability to convey contemporary spoken English in Dublin. Our study consists of a corpus analysis of three of his novels, where we are able to describe a number of key stylistic features which are distinctive of Howard’s representation of Irish English, including the variable use of such discourse markers as like, roysh and yeah + no, as well as the use of the intensifying so. INTRODUCTION The appearance of traces of orality in fiction has been thoroughly investigated in the field of stylistics, where literary dialects are a fertile ground for research. The use of dialect in fictional contexts is a tool often employed by writers to foreground the regional identity of characters. In such contexts, different voices are represented as a symbolic depiction of complex linguistic realities that contribute to characterisation. The term representation inevitably points us from the outset in the direction of the issues of realism and authenticity, while bringing to the fore the question about whether fictional dialogue/dialect should be judged on how closely they resemble real spoken discourse. What value does the representation of orality add to a novel, a film/TV script or a play? Is it a key element in the reception of the text by the audience/readership? Does it determine in any way the success of a text? In order to allow researchers to examine how closely a text represents spoken data, various models have been developed that help us measure the distance between spoken and written language, taking into consideration different factors. Jucker’s model (1998), for example, considers genuinely written data, that is, texts that are intended to be read and written representations of spoken language, which he subdivides into ‘retrospective’, ‘fictional’, and ‘prospective’. In this model, narrative fiction containing dialogues between characters, or dialogues embedded in first or third-person narration falls into the second category. Alternatively, Schneider’s model (2002) distributes written sources in linguistic studies that resemble spokenness more closely into five categories. The first and more ‘faithful’ category is ‘recorded texts’, which encompasses the sources that reflect speech that is immediately written down on site (Schneider 2002: 72). ‘Recalled’ and ‘imagined’ texts, which resemble reality less closely, are the second and third categories. This article, how- ever, is concerned with Schneider’s fourth and fifth categories: ‘observed’ and ‘invented’ ∗ Senior Lecturer at the University of Extremadura, Department of English University of Extremadura, Avda. Univer- sidad s/n, 10071 C´ aceres, Spain. E-mail: camador@unex.es ∗∗ Doctoral student at the University of Extremadura. E-mail: annterrazas@gmail.com C  2017 John Wiley & Sons Ltd