Skip to main content


Lexical Access in Bilingual Speech Production ‘’the hard problem’’: Evidence from Masked Priming of Picture Naming – Manal Alharbi

Past psycholinguistic research on bilingual word production investigating the manner of language selection process has adopted various methodologies, including the picture –word interference, language mixing/switching and phoneme monitoring paradigms, and produced conflicting results which might be attributed to task-related flaws (Costa et al. 2006; Kroll et al. 2010). Thus, in this research the masked priming paradigm in a picture naming task was used to investigate, for the first time, lexical access in Arabic-English bilinguals who use with different scripts. In experiment 1, subjects were asked to name cognate and non-cognate pictures that were preceded by a masked L1 prime word. Prime words were either related (i.e. the name of the picture in L1) or unrelated. Results showed a significant priming effect for cognates and non-cognates when preceded by related masked primes. This was interpreted as evidence that non-target lexical nodes were active and do not compete for selection. In experiment 2, subjects were asked to name non-cognate pictures that were preceded by masked prime words in L1. The prime words were either semantically related or unrelated. Results showed a semantic interference effect which was interpreted as evidence of lexical competition at the lexical level. A third experiment is on-going to investigate these contradictory findings.

The Phonological Effect of Iraqi Arabic as a Heritage Language on Cardiff English as a Majority and Native Language for Adult Heritage Speakers of Arabic in Cardiff, the UK – Hamed Aljemaily

This research is intended to investigate the phonological effect of Iraqi Arabic as a heritage language on the Cardiff English of relevant native speakers in the UK. It will attempt to find out any possible interference from Arabic as a non-dominant language in the production of English as a majority language. In other words, it will auditorily and acoustically test whether Iraqi Arabic as a minority language influences the production of English spoken in Cardiff, UK. To elaborate, the concentration will be on whether a peculiar accent is detected in their articulation of English segments based on accentedness ratings; and to look at and pinpoint the acoustic characteristics of some segments to determine any potential deviation on the segment level that might result in different phonetic properties in these segments. Furthermore, it will discuss the correlation between accentedness ratings and the acoustic measurements.

For the methodology, this data has been collected from two groups: an experimental group (namely, heritage speakers of Iraqi Arabic) and a control group (namely, monolinguals in Cardiff English). The implemented instruments are a reading material and a demographic questionnaire. Then, their audio recorded readings will be rated by ten monolinguals in Cardiff English. In addition, acoustic characteristics for a number of segments will be measured and compared between the two groups. The findings will present and discuss any possible overlap between these two languages focussing on the interference from Arabic in the articulation of English.

The Role of Digital Multilingual Patterns & Functions on Branding and Self-Branding: An Investigation of Practices by Saudi Users on Twitter – Reem Al Madani

The increasing number of internet users means that corporations and social media influencers build creative social media strategies and interactive posts to engage with their followers and encourage them to share the content related to them and their brand. The research project examines the digital communication of corporations, social media influencers, and ordinary users with a focus on the multilingual practices used to communicate with the existing and potential virtual audience on Twitter in Saudi Arabia. The study aims at investigating the use of the different forms of Modern Standard Arabic, Colloquial Arabic and English, in addition to the use of punctuation to express emotions and emoji in a corpus of more than 14,000 tweets from corporations, social media influencers, and ordinary users.

The poster will discuss the quantitative analysis of the data collected from the corporations’ accounts (9370 tweets). It will explain the different language varieties and functions that emerged from the data. Furthermore, the poster will showcase how the language choice plays a role in the function of the tweet.

A Cross-Linguistic Study of Metadiscourse Markers in English Academic Writing of Saudi EFL Students and UK Native Speakers of English – Nasser Alqahtani

The use of metadiscourse (MD) in academic writing is a concern for students due to its explanatory and rhetorical nature (Nash, 1992). MD markers are words (e.g. first and surprisingly) that guide readers in a text and show the stance of the writer. MD has two main functions: textual and interpersonal, which are based on Halliday’s (1973) macro-functions of language. On the textual level, MD guides the readers in the text to organize and interpret the content, and on the interpersonal level, MD turns the readers to the writer’s perceptions and attitudes making texts more reader-friendly (Crismore & Farnsworth 1989). Crucially, the use of MD markers varies from one group to another depending on the cultural and institutional contexts (Hyland, 2005; 2017). Therefore, specific practices are witnessed in Saud students’ writing as they come from a different society and culture (Alowayid, 2018). A few studies have looked at Saudi students’ use of discourse markers, and concluded that students underuse and misuse markers which result in low quality of their writing (Daif-Allah & Albesher, 2013; Alowayid, 2018). Most MD studies investigate MD based on a functional approach i.e. communicative function (Hyland, 2005) but this study combines the functional approach and the basic constituents of the clause (subject, predicate, etc.) to determine the functions of the markers syntactically and communicatively. Using a corpus of 30 MA dissertations in Applied Linguistics, this study explores the differences and similarities of MD usage between three groups of participants: Saudis in Saudi Arabia, Saudis in the UK, and UK native speakers of English. This research adds to the emerging L2 academic writing literature on the use of MD cross-culturally. The results can be used to develop materials for teaching in KSA and to see what affect students’ use of MD more the discipline or the institutional context (Saudi universities VS UK universities).

Positive and negative evaluative language: A comparative study on British news reports of the ISIS conflict in Iraq and Syria using Appraisal theory – Zeen Al-Rasheed

Evaluation language conveys attitude towards an individual, situation and other beings or objects. It is both subjective and exists within a societal value-system (Hunston 1994: 191). I draw on Martin and White’s Appraisal theory (2005) and some other studies such as: (Chiluwa and Ifukor 2015, Chen 2014, Julian 2011) in order to investigate whether British journalistic accounts of Iraq and Syria produce similar or different evaluation patterns and to find out the types of group identities.

The data I have chosen encompasses twelve British news reports from a range of newspapers. I divided them into two sets: six of the sets related to the ISIS conflict in Iraq and the other six related to the conflict in Syria. In the poster, the main point I demonstrate is the analysis outcomes regarding both sets of press reports. I show in a table the evaluative pattern results for some targets i.e., the appraised characters that are involved in the ISIS conflict. Based on these patterns, the similarities and differences among the targets have also been identified. Finally, the classification of some targets as good or bad group is also demonstrated. There are, however, two of the targets that I have not decided yet under which group they could be classified and therefore I have included two questions which are related to this issue of grouping.

The Non-arbitrariness of the Emotive Hijazi Non-Lexical Expressions – Mashael Assaadi

In everyday speech, Hijazi Arabic is full of non-lexical Expressions (NLEs) that constitute independent utterances, and realize the speaker’s emotional states. For example, wah! realizes surprise, jɪʕ! realizes disgust, etc. My study investigates the non-arbitrary relationship between Hijazi NLEs (HNLEs) and their emotive meanings by considering them as semiotic signs. In semiotics, a sign is anything which refers to something else. Signs communicate meanings. NLEs signal emotional states and hence are signs.  

Saussure (1916) states that signs are arbitrary; their meaning is not predictable from the form. However, he allows for some exceptions such as interjections. I argue that the NLEs are similar to interjections, as both of them are “spontaneous expressions of reality dictated …by natural forces” (ibid 69). In other words, they show the interaction between the socio-cultural and intrapersonal aspects of language (Wharton 2003; cf. Wierzbicka 1992). Like interjections, NLEs are inner expressive speech with unusual vocalizations depending on specific situational and socio-cultural contexts. Based on Hjelmslev (1963) and Halliday and Matthiessen (2014) framework, I argue that the emotive HNLEs are semiotic signs that contain an internal dynamic system of stratification between two orders of the Hijazi NLEs’ abstractions: the content plane of the Hijazi NLEs (i.e. the emotional meanings) and expression plane of the Hijazi NLEs (i.e. the phonetic form).

This study examines 34 HNLEs whose meanings have been checked through an open questionnaire that elicited responses from Hijazi speakers of different social variables. I classify their answers to explore the non-arbitrariness of those HNLEs. They have been mapped with Shaver et al.’s (2001) emotion classification. The data analysis shows that the HNLEs of every basic emotion are similar in their physical vocalizations. In other words, the content of a specific emotion has a meta-redundant relationship with the expression of the emotive Hijazi NLEs that realise this specific emotion, as they share some common vocalisations and emotional meanings. These common vocalisations show iconic representations, which in themselves are indexes for that emotional meaning. Specific feelings motivate us to shape our mouth in a specific way. We produce those NLEs which have unusual phonological structures to communicate these meanings (c.f. Darwin 1872: 83, 91).

Global Bilingual News: Same events, but variable perspectives – Banan Assiri

From Tunisia and rapidly across Egypt, Yemen, Bahrain, Libya and Syria, and recently in Sudan, Algeria and Iraq, the ‘quest for democracy’ movements that became popular in the media as the ‘Arab Spring’ attracted global media interest. Besides their production and reception variation (Clausen 2003), international news media may generate different emphases of the same event and assert different forms of narration to support particular political, ideological and social perspectives. In this poster, I describe how I employ both cross-linguistic corpus linguistics (Taylor, 2014; Vessey, 2013) and critical discourse analysis to discusses how reporting of the same events by the same news outlet in two languages may lead to a varied and contradictory representations of particular political and social actors and voices. This is of interest because it may problematize the notion that an individual news source has a coherent news story or stance and adds empirical support for the notion of ‘representation’ in the reporting of events (Hall, 1999). Also, I display how I apply News Values approach (Bednarek & Caple, 2014) joined with the voices analysis (Bednarek, 2016), to explore representation of political and social actors in the ‘Arab Spring’ in the Arabic- and English-language articles of two international news outlets, Aljazeera and the BBC. In both Aljazeera and the BBC, there was evidence of how discourse features change to index legitimacy in the representation of political and social actors’ voices in the ‘Arab Spring’ from contrasting cross-linguistic/cultural perspectives.

Using CADS to uncover how right-wing extremists generate an alternative rape culture in online narratives – Kate Barber

Right-wing extremism, in its many guises, is spreading and expanding internationally (Caiani & Kröll 2015, p 331), particularly through online platforms.  Linguistic analyses of far-right discourses have traditionally focused on nationalist rhetoric or racist and ethnoreligious-based invective (see, for example, Brindle 2016). Recently, academic focus has been on group identity within the far-right (Gray 2018, Lorenzo-Dus & Nouri 2020).  However, the explicit anti-feminist stance held by some far-right groups, and how this contributes to their collective identities and ideologies, remains underexplored.  Related to this, the growing men’s rights activist movement known as the Manosphere is appearing to move closer to the far right online (Romano 2016, Marwick & Caplan 2018).  The overall aim of this study is to highlight the extent to which ascribed and inhabited identities in both the Alternative Right (Alt-Right) and the Manosphere converge in the way they portray victims and perpetrators of sexual violence, particularly through the manipulation of narrative discourses.

The poster outlines findings from an ongoing corpus-assisted critical discourse analysis of blog posts from Manosphere sites. Following a brief overview of the data and the XML mark up of lexical and clausal elements of the posts, the poster examines two aspects of the study: 1) the organisation of narrative components in news narratives from Manosphere blogs; and 2) how identities shift between narrative components to generate a shared alternative rape culture which excuses sexual violence against women while promoting white male victimhood.  It is hoped that this study will contribute to research on the relationship between identities and narrative in extremist rhetoric and, eventually, to work on countering violent extremism online.

English language vocabulary pedagogy: Investigating the impact of explicit etymology and morphology teaching on children’s vocabulary development in schools in Wales – Ellen Bristow

The ability to cope with different levels of lexical complexity is fundamental to an individual’s success, both academically and in life post-education. To ensure individuals acquire the varying  levels, registers and a range of vocabulary necessary for this success, the Curriculum for Wales (2016) purports to expose children in Wales to the high-level and variety of vocabulary required. However, in Wales in 2019, the number of children reaching the expected standards in reading and writing by age eleven (end of Key Stage 2) decreased by 1.4% (Welsh Government 2019). Additionally, in 2019, the English language GCSE pass-rate in Wales dropped by 4.1%, with just 59.2% achieving an A*-C grade compared to 61% in 2018 and 63% in 2017 (Qualifications Wales 2019). This evidence suggests that there is a ‘gap’ between the vocabulary children should know and understand, compared to what they actually know and understand. Consequently, this doctoral project takes this vocabulary ‘gap’ as its central concept. Drawing upon theories of child vocabulary development (e.g. Clark 1993), derivational morphology acquisition (Tyler and Nagy 1989), English language pedagogical theory and practice (e.g. Quigley 2019; Murphy and Murphy 2019), and progression outcomes from the current and Draft Curriculum for Wales 2022 (2016; 2019), it examines whether integrating explicit etymology and morphology teaching into English language education could influence children’s ability to use and comprehend complex academic vocabulary.

A Methodological Approach to a Discursive Analysis of Judicial Activism – Débora Cabral

This PhD research aims to explore how judicial activism and judicial bias are linguistically enacted, and to what extent there is an institutional identity of impartiality in Judge Moro’s Lava Jato sentencing decisions. Operação Lava Jato (Operation Car Wash), a Brazilian Federal Police investigation, unveiled a network of corruption that included executives from major Brazilian companies and prominent politicians. These sentencing decisions will be examined under the light of corpus linguistics and van Dijk’s sociocognitive critical studies (2008; 2014; 2018) in order to contribute to the worldly discussion of the role of the judiciary in a democratic society. This poster presents specifically the methodological approach to analyse judicial activism and some of the next steps to analyse judicial bias and impartiality.  Judicial activism has been an ongoing discussion among law scholars (e.g. Baptista and Puerari 2017; Castilho 2018; Dias and Zaghlout 2016; Kmiec 2004; Nascimento 2011; Ribeiro 2018). In the linguistic field, however, no linguistic work on judicial activism has been found. In the Brazilian setting, there are studies on polyphony, i.e. the many voices, present in sentencing decisions (Catunda 2010, Valverde et. al 2013), argumentation in sentencing decisions (Pistori and Banks-Leite 2010), stereotypes and subjectivity (Cabral and Conceição 2017), and how social actors in the penal court context are represented (Fuzer 2010), but there is no study on judicial activism. For that reason, before examining Judge Moro’s Lava Jato sentencing decision corpus, it will be necessary to identify the types of linguistic features that might characterise activism, and then, more specifically, judicial activism, and only then analyse Judge Moro’s sentencing decisions.

An Empirical Investigation into the Nature and Degree of Nominality – Alex Carr

While it is widely accepted that nouns and verbs constitute two separate lexical classes, this distinction has been shown to be unsatisfactory due to variably applied criteria which inconsistently mix morphology, syntax and semantics (Lyons 1977:423). For example, one problem is that semantic distinctions such as event vs object are not reliable, e.g. the noun ‘fire’ behaves verbally, like a deverbal noun, but this meaning is not inherited from a verb (Vendler 1967:141). Hanks (2013:73) suggests that lexical items, in isolation, do not possess inherent meaning, but “meaning potential”, which is activated when placed in a context. In the field of Lexical Semantics, the analysis of word meaning from the ‘Semasiological’ perspective provides a beneficial heuristic to view the various meanings a nominal can express. By exploring the multidimensional interconnections of meaning within lexical items, we can examine how such nominals can express both event and object semantics. Nevertheless, while this analysis has allowed us to explore the semantics of nominals at a level that cannot be attained from only looking at classifications of lexical class, it does not explicitly provide us with empirical data on how the semantics of these nominals generally behave. Therefore, the aims of this research are threefold: (1) to determine how the nature and degree of nominality can be evaluated, (2) to determine how object, state, and event meaning come to be expressed in nominal forms and (3) to examine the relationship between the syntactic behaviour and the semantic properties of underived event nominals (UENs).

A contrastive study into the lexico-grammatical reactances of two intransitive verbal categories – Lucy Chrispin

Within the theory of Systemic Functional Linguistics (SFL), Halliday (1994) proposes six process types that each represents a different experience. These process types are concerned with the transitivity system, and are formed from the lexico-grammatical interface of the verb and surrounding participants and circumstances. Whilst some instances of language fit neatly into a specific process type category, others are not always obvious, and this can lead to discrepancies in process type analysis (O’Donnell et al. 2009). In particular, there lacks reliable lexio-grammatical reactances to distinguish behavioural processes from material processes that are intransitive, i.e. neither take a that-complement, they both favour the present progressive tense, and behaviourals are typically intransitive. This raises the question as to why these experiences are theoretically outlined in two separate categories. Thus, this paper seeks to test the theoretical claims about both categories using empirical data, and in doing so, attempts to identify any lexico-grammatical reactances that can be used to differentiate behavioural processes from intransitive material processes. The data includes a sample of five verbs that typically occur in behavioural processes and five verbs that typically occur in intransitive actions, using the Corpus of Contemporary English (Davies 2008-). The paper highlights the value in using Hanks’ (2004) Corpus Pattern Analysis (CPA) combined with aspectual analysis (Van Rompaey 2013, pp.181–219) to reveal similarities in four lexico-grammatical reactances – animacy, intransitivity, -that complements, and grammatical aspect – with the major difference lying in the lexical aspect.

Determining the impact of Syntactic Intelligibility on the Regognition of the Emotional Speech Signal – Matthew Coombes

The result was a survey which invited participants to listen to a number of voices speaking short sentences, and to then rate their affective response through a series of likert-based scales. As before, three sets of spoken-word stimuli were produced – the first being recognised lexically and grammatically as English, with the second containing reduced levels of recognisable grammar and syntax – with the third set (constructed language) reduced to minimal values, thus meaning English was in no way distinguishable from the spoken words.

Rather than approaching an international audience (as was the case for the pilot study), it was decided to target a population within a given area – later to be decided as the county of Cornwall. This would ensure that bias between dialect, accent and other such nuances would be kept to a minimum. The data from this survey would then be used later when subsequent studies would be targeting more metropolitan areas.

The purpose of accessing this data is to represent a profile of aesthetic preferences to vocal acoustics, and how this differs both nationally and internationally with respect to the aforementioned factors relating to geolocalisation.

Keepin’ it real: Styling authenticity on Twitter disinformation accounts – Aurora Goodwin

Disinformation is currently one of the biggest threats to global democracy (World Economic Forum 2016), causing governments to create regulations to censor and control distributions of ‘fake news’ (West 2017). One of the most prominent disinformation campaigns was that led by the Internet Research Agency during the US Presidential Election of 2016, leading to the indictment of 13 Russian nationals on the grounds of ‘defrauding the United States’ (US Department of Justice 2018).

Using a pre-existing data set compiled of a total of 3837 accounts identified as belonging to the Internet Research Agency, I aim to combine qualitative and quantitative methods to investigate authentication processes carried out by some of the most followed accounts. The analytical chapter currently in progress is aimed at investigating how disinformation accounts style their profile meta-data to authenticate performances of identity.

Metadata is broadly used to refer to ‘data about data’ (Foulonneau and Riley 2008:3). Within social media research, specifically on Twitter, metadata has been defined as ‘characteristics of a Twitter profile that (are independent of tweet content)’ (Morgan-Lopez et al 2017). I focus specifically on usernames and handles, and hope to expand this to profile pictures and bios in the future.

This poster will focus on the progress of my current chapter, specifically my analysis of usernames and handles, and their contribution towards authenticating processes.

What’s So Special About The Circus? (And Who Says So) –Katharine Kavanagh

This research is a corpus-assisted critical discourse study into the values attributed to the experience of attending circus performances. Underlying concepts of value are seen to materialise through cognitive stances that are realised in texts as evaluation (Thompson and Alba-Juez, 2014:10; Bednarek and Caple, 2017:78). Evaluation is also a key registerial feature of promotional genres, both those framed as marketing and those framed as reviews —otherwise known as ‘interested’ and ‘disinterested’ genres (Shaw, 2009). In the circus field, my previous research has found that reviews readily available to the British public only represented a limited variety of the values that can be attributed to the experience of attending circus productions (and did so in a non-dialogic, authoritative fashion, occluding the possibility of alternatives). It follows that there are experiential values held by members of UK circus audiences that are not publicly represented. To discover the extent to which this is the case, and the way any such discrepancies manifest, my study combines Audience Research and Corpus Linguistics methods to compare the evaluation realised in three subcorpora: audience interview transcripts; circus reviews; and corresponding publicity blurbs. The various values represented will be identified by triangulating key semantic domain analysis (Rayson et al., 2004) with APPRAISAL analysis (Martin and White, 2005). The keyness analysis will use the full circus evaluation corpus as reference, while each subcorpus will be proportionately sampled, with the samples coded for APPRAISAL features using UAM CorpusTool (O’Donnell, 2008). This poster presents the rationale for this research and the proposed methodology.

Morphological Regularities and Patterns in English Word Formation – Kateryna Krykoniuk

The poster provides an overview of the thesis entitled Morphological Patterns and Regularities in English Word Formation. It, first, centers around the aim, the methodology and theoretical problems substantiated in the research. As follows from the dissertation title, its main objectiveis to identify fundamental patterns, relations and regularities that determine how morphemes are organized in English words. Further, different methodological toolsaredeployed in this research. Specifically, a sample of lexemes (40,000) was formalized with the help of formal morphological analysis, which transforms lexemes into formulas containing morphological information. Then, the morphological formulae were classified and organized using macros written in the VB, which resulted in the formation of a morphological meta-corpus—a set of formally represented morphological patterns of English words. Lastly, various statistical analyses are used to explore various aspects of this meta-corpus (e.g. Principle Component Analysis, regression modelling, conditional inference trees and forests). In addition, theoretical problemssubstantiated in this thesis concern (a) constraints of suffix ordering and the interaction between affixes and word bases, (b) the impact of type and token frequency of morphemes and of morphological patterns on English word formation grammar, (c) such typological characteristics of English as isolation, agglutination and fusion, and (d) the relatively newly-established concepts of global and potential productivity (Baayen 1991). The second part of the poster describes the structure of the thesis by looking at the dissertation chapters. Finally, at the bottom of the poster, two questions are raised for the purpose of the feedback.

Victim Voices: Analysing Child Discourse in Online Sexual Grooming Interactions – Ruth Mullineux-Morgan

Online Child Sexual Grooming (OCSG) is an internet-enabled communicative process of entrapment in which an adult uses language and other semiotic modes (e.g. images) to lure a minor into taking part in sexual activities online and, at times, also offline (Lorenzo-Dus et al 2020:16). Research into the language used in OCSG has proposed an Online Grooming Discourse Model (OGDM) which reveals that offenders regularly deploy a range of processes (Deceptive Trust Development, Isolation, Compliance Testing and Sexual Gratification) to carry out sophisticated relational work to create interpersonal closeness with their target victims (Lorenzo-Dus et al 2016). What we currently lack is an understanding of how child victims of OCSG interpret offenders’ strategic use of language.

This poster outlines the proposed approach aimed at filling this important gap in knowledge. Data comes from three hitherto un-examined sources (i) children’s accounts of their experiences of OCSG in a counselling context via a national helpline; (ii) police data of social media interactions between children and groomers; and (iii) focus groups with child protection and helpline counselling professionals.

Early findings from an analysis of the child’s discourse about the groomer’s communication and behaviour are presented. These reveal a dissonance with the OGDM, with 79.4% of children’s utterances within the Compliance Testing process seeming to fall outside previously identified sub-strategies. Instead, children refer to threats, blackmail and harassment, appearing therefore, to indicate the presence of communicative coercion (Chiang et al 2018; Schneevogt et al 2018; Chiang & Grant 2019). In order to probe this finding further this research will draw on im-politeness theory (Brown & Levinson, 1987; Culpeper 2011). Further, the child’s evaluation (feelings and opinions) of their experiences of OCSG will be analysed drawing on the modified AFFECT subsystem of Appraisal theory (Martin & White, 2005; Bednarek, 2008) proposed by Benitez-Castro & Hidalgo-Tenorio (2019).  

Automatic grammar induction from free text using insights from Cognitive Grammar – Vigneshwaran Muralidaran

Identifying the grammatical structure of a sentence is useful in many Natural Language Processing applications such as Document Summarisation, Discourse Analysis, Language Prediction and Machine Translation. With the availability of syntactic treebanks, supervised parsers have been developed successfully for many major languages. However for minority languages with low resources, this poses a challenge. My research aims to find a linguistically motivated approach that automatically induces grammatical structures from free sentences with very little or no annotation. Although primarily developed for Welsh (and English), this work can be readily adapted to different languages including low-resource minority languages. We draw the basic outlook of linguistic analysis from Cognitive Grammar. Our approach identifies grammatical patterns from a free text by recognising domain-independent, general, cognitive patterns of local coherence that occur in natural language utterances. The implementation is also motivated to reflect some of the general psycholinguistic properties of parsing by humans – such as incrementality, connectedness and expectation. As there are multiple valid ways to assemble component schemas into a complex, there is no one gold standard to compare the output against. Therefore we evaluated the system by generating new sentences using the component schemas as templates and asked human volunteers rated them on a scale of 1 to 5. Using Wordnet information along with Parts Of Speech tags resulted in better prediction and improved the rating of the sentences generated for evaluation in English. We have to yet conduct the volunteer evaluation for Welsh.

Following a long Summer of compartmentalising, an updated thesis structure based on the pilot studies was completed during my second year. A number of improvements relating to the human-computer interaction via the interface design. along with some of the wording and the questions pertaining to the necessary information required from the participants were made in retrospect to the pilot.

The Language Of Student Complaints Procedures In The UK Higher Education Sector – Lisa Pomfrett

Regulatory, policy and guidance documents are a common feature of the modern Higher Education sector.  Yet, whilst research on the language of the classroom and academia is growing, there is a distinct lack of research on the language of these documents.  My research will add to existing knowledge of the language of non-academic text in the sector through looking at these documents as both a tool for the reader to use for guidance and as a mechanism to create / reinforce internal relationships within the sector. 

My research will explore written communication in the Higher Education sector surrounding student complaints procedures.  Using a genre-based approach and corpus linguistic methodologies, relevant regulations, policies, student guidance and staff guidance documents will be analysed to determine how obligation and commitment to act are communicated; if there is evidence that documents can be considered individual genres and representation of hierarchical / power relationships within the documents.  The research is embedded in practices of the modern Higher Education sector and data used for this research aims to reflect this through looking at all users of the documents (staff and students) as well as selected missions groups that represent two distinct types of Higher Education institutions in the current sector (Russell Group and Post-1992).

‘All incriminatory information in this work is written “in character” and must not be confused with an actual plan, or strategy to attempt to harm any individuals or infrastructure’:  Virtual responsibility in criminal narratives – Emily Powell

The positioning of the self as similar or different to a past or future self (Bamberg 2012) is particularly relevant to criminal narratives because of the possibility that someone who intends to harm others may enable themselves to act by compartmentalising different parts of themselves in some way (Cohen 2001). The creation of a virtual self within such narratives may therefore enable an offender to preserve their own moral identity by viewing their future acts as separate to their current identities as good people (Maruna 2004), or conversely, may allow the killer to rehearse their agency and their role as killers through their narrative using hypothetical situations (Brockmeier 2009; Bruner 2004).

This study uses a corpus linguistic approach to diachronically analyse stylistic changes in a corpus of texts written by mass murderers (Anders Breivik, Elliot Rodger, Dylan Klebold, and Eric Harris), and explores the way in which they navigate their moral agency by presenting themselves as similar or different to their hypothetical or virtual selves, and linguistically blurring the lines between virtual and real-life killing.

The vocalic inventory of late Old Northumbrian – Elisa Ramíres Pérez

Unlike Present-Day English, Old English (ca. 7th c. to 1150) is considered to be a rather phonemic language, in the sense that there existed a more robust sound-letter correspondence. While the reconstruction of the Old English consonantal system has presented challenges of its own, it is the reconstruction of the vocalic system of Old English which has proven most problematic (Minkova 2014: 175-176). According to the authorities in the field of diachronic phonology, the Old English vocalic inventory contained short and long vowels as well as short and long diphthongs (Campbell 1959; Lass 1994; Hogg 2011; Minkova 2014). In addition to these, there also existed a number of vocalic digraphs, that is, combinations of two vowels, which, despite their similarity in spelling, did not behave or develop phonologically like diphthongs did. The main complication for my PhD project, which studies the weak verbal system in the Northumbrian dialect of Old English (10th c.), is that the phonological accounts of Old English are majorly based on the semi-standardised West-Saxon dialect. It is known that these two dialects are considerably different. Most crucially for my project, Cole (2019) demonstrated that Northumbrian spelling reflects changes in pronunciation more faithfully than West-Saxon does. Similarly, Stenbrenden (2016) convincingly argued that the development and orthographic representation of certain Old English long vowels differed in northern dialects from more southern ones. Based on these recent discoveries, my poster will outline how I intend to interpret and classify vowels as presented in Northumbrian texts.

Question types and constraint in police-victim interaction: A case study – Kate Steel

This presentation examines the questions asked by first-response police officers who are attending a reported domestic abuse incident. There is a wealth of research on questioning in investigative interviews (Oxburgh et al. 2010), including work which examines particular strategies in relation to witnesses’ vulnerability (e.g. Aldridge and Luchjenbroers 2008). First-response call-outs also centre on the pursuit of institutional goals through questioning; eliciting the reported victim’s first account of ‘what happened’ is the primary means by which officers determine if an offence has occurred and ascertain the evidentially relevant details. Yet call-outs differ from investigative interviews in key aspects, including the increased potential for urgency and distress, and the lack of official guidance around what officers can say (cf. MoJ 2002). Despite the centrality of questions within this unique speech context, there has been no previous empirical research in this vein, presumably due to the complexities around researcher access.  

The case study presented forms part of a wider investigation into officers’ control of victims’ voices and their interactional spaces during first-response call-outs. Data extracted from body-worn video footage is analysed to address the following questions: Which question types are employed by the officers? To what degree do these questions constrain the scope of the victim’s responses? Analysis is informed by Newbury and Johnson’s (2006) typology of information- and confirmation-seeking questions. The presentation will demonstrate how officers’ selection of certain question types restrict the victim’s contributions, with implications in relation to her vulnerability and the reliability of the evidence produced.

Identity performance of older Greek-Cypriot Facebook users: Communicative Functions AnalysisStephanie Tilliridou

This study investigates how older Facebook users do identity work on their Facebook walls. The research question to be examined is: how do older Facebook users use semiotic and linguistic features to project different online identities on Facebook wall posts.  The data collection methods used include: 1) observing the participants’ activity on Facebook and 2) collecting and recording their Facebook wall posts in the form of screen-shots during a 6-month period. 2842 posts were collected during a period of 6 months from 13 older Facebook users. Even though ‘the age at which people are considered “older” is variable’ (Van House 2015, n.p), Facebook older users are defined in this study as people who are aged over 45 years old.

This analysis chapter focuses on the data analysis of Facebook wall posts based on their communicative function in order to identify the purposes these wall posts achieve and what identity work is done through achieving these purposes. The analysis was partly directed from previous literature (e.g Lee, 2011; Tagg et. al., 2017) using content analysis. Lee’s (2011) classification was used as a starting point for the categorisation of communicative functions, while new categories derived from the data in question.  By taking into account the types of events and activities the participants selected for sharing with their Facebook friends, the ways in which the participants present themselves to their social network could be inferred and, thus, the range of identities they deem relevant for this type of interaction.

Identification and Analysis of Visual Irony in Political Cartoons of the Algerian Civil War (1992-2002) – Sabrina Toumi

My PhD research looks at metaphor, metonymy and irony in political cartoons depicting the Algerian civil war and the high state censorship imposed on the independent press throughout the 1990s (Labter, 1995; Evans and Philips, 2007). The 540 cartoons in my data set were published in El Khabar, El Watan, and Liberté newspapers between January 1992 and December 2002. I examine the rhetorical figures both in isolation and regarding how they interact in many instances. The present poster is concerned specifically with the analysis of visual irony in the cartoons. It reports on the results as well as addressing some of the challenges encountered during my pilot analysis of a hundred cartoons which I selected randomly from the larger data set. Adopting the echo-mention view of irony (Sperber and Wilson, 1981), my analysis aimed to identify the occurrences, the properties, and the functions of visual irony in the cartoons.  I also drew on Clift’s (1999) notion of irony as ‘framed evaluation’ which serves as a vehicle of criticism towards the object/victim of irony. These views are coupled with Scott’s (2004) distinction between word-based and wordless types of visual irony. While irony is found to be a potent figure in the analysed cartoons, it is a very complex figure to capture.  

The Acquisition of Sociolinguistic Competence in a Welsh Immersion Context – Katharine Young

Previous research in Canada has shown that students in French immersion classrooms often struggle to acquire the range of styles available to L1 speakers and so, it could be said, do not reach full sociolinguistic competence (Mougeon et al., 2010). Welsh-medium education differs insofar as pupils from Welsh-speaking and non-Welsh-speaking homes are taught together and that the proportion of pupils from different language backgrounds differs between areas. Despite work on the acquisition of grammatical features (e.g. Binks & Thomas, 2016) and phonological variation (e.g. Mennen, Kelly, Mayr, & Morris, 2020) in the speech of pupils in Welsh-medium education, little is known about;

 a) the stylistic repertoire of pupils from Welsh-speaking homes;

(b) the extent to which pupils from non-Welsh-speaking homes acquire this repertoire; and

(c) whether the sociolinguistic competence of those from non-Welsh-speaking homes differs between Welsh-dominant and English-dominant areas.

Previous work (Young, 2019) on stylistic variation in Welsh classrooms has shown that, according to teachers, the range of variation between formal and informal variants is highly dependent on the feature and changes in register (spoken vs written, classroom vs outside the classroom). However, this data was drawn from teacher reports across types of Welsh teaching (Welsh as a First Language and Welsh as a Second Language), and the current project will focus on stylistic variation produced by pupils of Welsh as a First Language only (where home language backgrounds are mixed), in a number of registers, with a view to understanding the constraints on the acquisition of sociolinguistic competence in Welsh-immersion settings.