Discussion around the importance and prevalence of multiword expressions in the lexicon and the teaching of vocabulary has existed for a number of years in applied linguistics (e. g. lrujo, 1986; Pawley and Syder, 1983; Sinclair, 1987; Wray, 2002). While there seems to be a general agreement among scholars that formulaic language should feature in language learning and, perhaps to a lesser extent, language testing, there appears to be rather less agreement when it comes to how to select and/or prioritize specific items for inclusion. One criterion for selection which has been used often for vocabulary items of single words is frequency (i.e. how relatively common a word is), data for which can be consulted using various frequency lists that have long existed and are in the public domain, such as the General Service List (West, 1953). However, to date, no list of formulaic language that could be considered comparable to the General Service List in terms of intended use and relevance to language instruction has been attempted. The work presented in the present thesis aims to address this lack. The thesis first presents the need for such a list, and then describes the methodology employed by the researcher to ultimately produce a frequency-informed and pedagogically-relevant list of multiword expressions that can be used in conjunction with existing lists single orthographic words to help inform such instruments of L2 pedagogy as language textbooks and language tests, entitled the PHRASal Expressions List, or PHRASE List. To that end, two projects are also presented in the thesis which exemplify ways in which the list may be usefully employed. The first is a research validation exercise carried out in collaboration with the English Profile project in order to compare the phraseological component of the English Profile Wordlist to the expressions in the PHRASE List. The second project presents the development and validation of a kind of vocabulary test that samples from the PHRASE List, and which is intended to be used to supplement knowledge assessed in existing tests of single orthographic words, such as the Vocabulary Size Test (Nation & Beglar, 2007).

Despite the conceptual importance of investigating the social context(s) in which second-language (L2) learning and use take place, the decade-old “social turn” within the field of second-language acquisition (SLA) has yet to produce a “parsimonious system of valid and generalizable parameters to describe contextual characteristics” (Dörnyei, 2009a, p. 238). Accordingly, investigating social network structure has recently been suggested as a general approach to examining the link between person and environment (Beckner, et al., 2009). In the current thesis, I offer a network approach in which second-language (L2) learning and use is regarded both as purposeful and as constrained by one’s social relationships. Subsequently, in a first-of-its-kind study within SLA, I apply social network analysis – a diverse array of formally-defined measures of social position and other socio-structural features – to conceptualize and empirically test the relationship between social structure and the willingness to communicate in the L2 (L2 WTC), defined as the “readiness to enter into discourse at a particular time with a specific person or persons, using a L2” (MacIntyre, et al., 1998, p. 547). In a study of Chinese-speaking international students at a British university, trait-like L2 WTC is found to predict cross-cultural adjustment, suggesting the role of actively engaging with one’s new cultural surroundings in establishing an adaptive person-environment fit. Subsequently, in the first study to apply modern, graph-theoretic notions of social position to a network of L2 learners, a significant relationship is found between various notions of structural position among a network of international English-for-Academic-Purposes students, and dispositional L2 WTC. Overall, the results support L2 WTC as both purposeful and constrained, learned from one’s past interactions, yet pushing the individual to take advantage of opportunities to communicate in the L2. Implications, limitations, and future directions of a social network approach to L2 learning and use are also discussed.

Language Learning has utilized technology for decades, and while world-wide social dynamics place more demands for language learning, there has not been a widespread use of a specific technology as the dominant medium for language learning. In the meanwhile, Virtual Worlds technology emerged during the last two decades as an immersive technology that offers an online representation of reality, allowing user interaction with the surrounding environment including objects and other users through Internet-enabled desktop personal computers. Since their introduction, Virtual Worlds have grown in popularity, and are now utilized by a large online community as social and gaming environments. Over two decades of research have shown the potential of Virtual Worlds for learning in various fields, but very few empirical studies have been dedicated to explore Virtual Worlds for language learning. The focus of this PhD research project is to explore the potential of the Virtual World Second Life in enabling effective language learning. The research question is as follows: ‘Could Virtual Worlds offer a suitable language learning environment, similar or better than that offered by traditional media of language learning?’ Towards answering that question, a pilot and two studies were conducted in 2007, 2008 and 2009 respectively. Arabic language classes were delivered to groups of language learners in the UK using different media of language learning: a face-to-face (f-to-f) classroom, a videoconferencing (VC) classroom, and a Virtual World (VW) classroom. The language learning quality outcomes along with student attitudes were assessed through a comparative analysis between the three media, involving attitude surveys, interviews, assessments of learning outcomes, and the critical incident method applied to video recordings. Due to several limitations, the effectiveness of the VW medium in enhancing the quality of the language learning experience was found lacking in the light of data collected and analyzed. A set of conditions and recommendations is therefore described to better utilize VWs for language learning.

This study reports on a corpus analysis of samples of online and face-to-face intercultural communication among a group of British and Taiwanese adolescents, with the aim of exploring the particular lexical, grammatical and discourse features of the online and spoken discourse from three perspectives: a keyness approach, a discourse analytical perspective and a multi-word sequence perspective. Keyness approach brings together three levels of keyness analysis: keywords, semantic domains and parts-of-speech, and further highlights those linguistic features that deserve particular attention. Furthermore, a discourse analytical approach adds greater detail and depth of description of language patterning by examining the particular linguistic features in context. Such findings that pertain to discourse and pragmatic functions in context are not likely to be made when only keyness is examined. The third approach of this thesis focuses on recurrent multi-word sequences, paying particular attention to their discourse functions in online and spoken settings. It is evident that multi-word sequences often perform systematic discourse functions, even though they do not usually constitute complete grammatical or idiomatic structures. The approach also examines the developmental perspectives of multi-word sequences, showing that intercultural contact with native speakers of English fosters the longitudinal development of the use of sequences by the Taiwanese learners. The method here, which focuses on naturally occurring language output, diminishes the effects of the artificial contexts often created in language testing settings. In light of the potential significance of the research to EFL pedagogy, the thesis further reports on the extent to which EFL textbooks used in Taiwan represent the particular linguistic features identified in authentic intercultural communication. The research findings demonstrate the pedagogical merit of the analyses of the three perspectives and thus help in the design of courses for adolescent intercultural interaction in both online and face-to-face settings.

This study develops and proposes a research approach founded on linking two novel theoretical frameworks for analysing and explaining relational strategies at intercultural workplace meetings, namely, the Multi-Level Model (MLM) and Cultural Self Perception (CSP). The approach is tested through a case study of a problematic meeting in an intercultural workplace in China. At this meeting, the Chinese and western expatriate participants both carefully adopted a range of relational strategies for problem talk. However, the interactions still evolved into a severe confrontation between the Chinese and expatriate participants. The cultural differences in relational strategies are explored on three levels, namely, tum-taking, speech act and situational context. The analysis indicates that while the relational strategies of the expatriate speakers can only be found on the speech act level, those of the Chinese speakers can be found on all of the three levels. Even on the speech act level, there are significant differences between the Chinese and expatriate speakers. It is argued that the communicative breakdown was caused by clashes of expectations of relational strategies for problem talk at meetings, and such differences can be explained by drawing on the framework of CSP. The investigation in this case study thus demonstrates that the combination of the frameworks of MLM and CSP can facilitate both systematic analysis of interactive strategies at different levels and in-depth understanding of the cultural roots for the choices of relational strategies at intercultural workplace meetings.

The practice of Teaching English to Speakers of Other Languages (TESOL) in the British context has evolved to a point where varied and complex patterns of classroom interaction have become the norm, and where teacher independent activities have become an important vehicle for language learning. In this climate, there is a need for novice English Language Teachers to adopt a major role as managers of learning. Whilst there has been much emphasis in recent Second Language Acquisition Research on the relationship between activity type and output from language learners, there have been fewer studies on the relationship between activity based pedagogy and teacher- generated language output. Using transcribed audio recordings from twenty-two entire lessons conducted by novice teachers, together with data from interviews and stimulated recall-based assignments, the author investigates aspects of lesson staging, classroom interaction and teacher role, as they are manifested through the language that the teachers use in their classes. Three types of teacher-generated language are identified. From these, it is suggested that the functions of structuring and rapport-enhancing have significance in lesson stages involving the setting up of teacher-independent activities. The complexity of the relationship between these two functions, seen in the context of entire lessons, reveals individual differences amongst the novice teacher subjects. implications of this research for Teacher Education are discussed, and the thesis ends with practical suggestions relevant to the content of Language Teacher Education Courses.

Learner autonomy is widely recognised as a desirable goal in tertiary education as it is found to comply with learner-centred approaches and enable students to pursue life-long learning (Sinclair, 2000a; Ciekanski, 2007). In language teaching and learning literature, it has become the main interest of researchers and practitioners who believe that learner autonomy can enhance students’ chance for success in learning a language. A great amount of research has been done to investigate various ways to foster learner autonomy in language teaching around the world (e.g., Benson, 2001; Breeze, 2002; Chan, 2001; Cotterall, 1995; Dam, 1995; Jing, 2006; Lo, 2010; L.C.T. Nguyen and Gu, 2013). However, learner autonomy is still widely considered a ‘western’ concept and much of the research has either been conducted in a western context or based on the western view of learner autonomy (Pierson, 1996, Sinclair, 2000a; Chan, 2001). This research aimed to gain more understanding of the development of learner autonomy in English language learning among students at a private university in Hochiminh city, Vietnam. The study has revealed that the major perception of learner autonomy in this Vietnamese context relates to ‘taking the initiative’ in learning, especially in selfstudy. The type of learner autonomy, as understood and practised by students in the context of Vietnamese tertiary education, has been argued to have the characteristics of Littlewood’s (1999) reactive autonomy. This finding lends itself to the application of Sinclair’s (2000a) teacher-guided/learner-decided approach to promoting learner autonomy. In other words, an integrated learner training programme (ILTP), which gradually developed students’ capacity to take more control in the learning process by providing them with metacognitive strategies for learning management, raising their awareness of themselves as learners and of the learning context, and encouraging them to explore the English language and its learning strategies, was perceived to foster the students’ willingness and enhance their ability to take the initiative in learning and create a habit of engaging more in self-directed learning. This study has also identified certain obstacles to promoting learner autonomy in Vietnam. In particular, the exam-oriented educational context poses significant challenges to both teachers and students in their efforts to promote autonomous learning. These difficulties include time constraints and a stringent syllabus. In addition to the contextual constraints, the large power distance between teachers and students in Vietnamese culture was also suggested to be a factor in hindering learner autonomy because it results in teacher reliance and an authoritarian view of the roles of teachers in the language classroom. This cultural trait, combined with the contextual constraints, seems to discourage teachers from giving students more control in the classroom and, at the same time, inhibits students from taking such control.

This study offers a contribution to the discourses surrounding 'silence', arguing that the transformation of the meaning of silence as it shifts between contexts depends upon what it is being defined in opposition to, and that in each case, what is posited in the space marked by silence is of central importance to the discourse surrounding this context. Aware of the interdisciplinary engagements with silence, this thesis presupposes that silence is not 'nothing', and that the question of agency is central to the distinctions between silences. Drawing on a number of theoretical perspectives pertinent to each context, this thesis proceeds by engaging with silence as it is featured in discourses surrounding animals, trauma, secrecy, and listening. These theoretical perspectives are explored also through a number of cultural texts - creative nonfiction, short stories, film, poetry, and also testimony. These case studies are not only illustrative, but also offer further perspectives on each context, and the meaning generated for silence. Unlike most other engagements with silence, this thesis not only takes the definition of silence to be unstable and changeable, but also confronts the question of why 'silence' is used in these discourses, positing that it is its association with space that is being drawn upon across these contexts. This thesis argues that it is because 'silence' comes to be figured as a creator of space, what is at issue in these contexts is what is conceived of as being in this space of silence - Otherness, isolation, individuality, intersubjectivity.

This thesis addresses a largely unexplored territory in second language motivation research. Its focus was to account for unusual persistence, effort and success among learners acquiring additional language(s) in conjunction with a sacred or special text. More specifically, the project sought to identify the factors, mechanisms and conditions that contributed to exceptionally high motivation for SLA. The guiding theoretical frameworks were Ushioda's Person-in Context relational theory and Dornyei's L2 Motivational Self System. A qualitative research approach was employed which included a series of in-depth interviews with 20 highly successful language learners. The findings confirmed both Ushioda's and Dornyei's theories. The Person-in-Context relational theory was helpful in understanding the significant role of the wider socio-religious context in which learners lived their lives. In addition, the three core components of the L2 Motivational Self System (Ideal L2 Self, Ought-to L2 Self and the language learning environment) shed light on these important components in the informants' learning trajectories. The research identified an ideal self in the form of a spiritual vision core to the informants' identities. In addition, the L2 vision was identified as inseparable corollary that enabled informants to achieve their spiritual vision. The third key component in the motivational matrix was the sacred text. However, beyond ideal selves, participants envisioned a shared ideal future grounded in imagined faith communities with concerns for broader societal well-being. The shared ideal future took form in the informants' imagination, not only within their lifetime, but beyond, extending to an 'ultimate' environment. The thesis includes implications for motivational researchers, language teachers and learners.

In this thesis, systemic functional linguistics’ long-assumed ‘context-metafunction hook-up’ hypothesis is subjected to its first large-scale, data-driven exploration. The claims embodied in the ‘context-metafunction hook-up’ hypothesis (henceforth CMHH) concern the relationship between language and context. Viewed as a set of relationships modelled with systemic primacy, linguistic phenomena group into three metafunctional sorts according to systemic functional linguists. The CMHH claims that these three metafunctional groupings correspond to three parameters of semiotic context such that they share a realisational relationship. The CMHH is one of the assumed strengths of the theory of systemic functional linguistics (henceforth SFL). Yet, despite its centrality to wider SFL research, ventures to test it on large-scale with naturally occurring language data are notable by their absence in SFL work. This project takes a step in the direction of filling the aforementioned void. Adopting Martin’s model of the contextual mode parameter as a starting point, the project proceeds on the assumption that if SFL’s CMHH is predictively sound, variation in ‘mode of discourse’ should correlate with variation in the occurrence of ellipsis in text. Assembling four different sub-corpora of natural language data varied in their contextual mode values following Martin – but otherwise in contextual identity – cases of ellipsis are coded along several variables. Statistical calculations are conducted on the results of this analysis. These calculations allow for detailed cross corpora comparisons which in turn allow for conclusions relative to the central research question to be drawn. The results suggest support for the CMHH at a broad level of generality. The most significant results in this regard are: (i) ellipsis is found to be more frequent the more ancillary a text’s context is; and (ii) the more ancillary a text’s context, the greater proportion of its instances of ellipsis are of the situational, rather than textual, type.