It is widely reported that second language (L2) speakers of English diverge from native speakers in their use of articles (the, a, 0) in two ways: they omit articles where they are required, and they assign interpretations to articles that are not those assigned by native speakers (Huebner, 1985; Ionin et al., 2004; Lardiere, 2004; 2005; Parrish, 1987; Robertson, 2000; Thomas, 1989; White, 2003a). Many of these studies have focused on speakers whose Us s lack articles (Korean, Russian, Japanese, Turkish). Within the framework of the Full Transfer/Full Access hypothesis about L2 acquisition (Schwartz and Sprouse, 1996), a number of proposals for explaining this divergence have emerged: articles are omitted because learners have difficulty mapping abstract syntactic representations into phonological forms (the Missing Surface Inflection Hypothesis-White, 2003a); learners assign non-target interpretations to articles because they are fluctuating between the definite and specific values of an article choice parameter (the Fluctuation Hypothesis-Ionin et al., 2004), or they have difficulty with `feature assembly' in the L2 (Hawkins et al., 2006; Lardiere, 2005). The predictions for speakers of Us that have articles that encode definiteness appear to be that these speakers will show much less divergence when they acquire English, although there is currently little evidence relating to such speakers. In this thesis, existing hypotheses about divergence in the use of English articles by non native speakers are tested in the context of LI speakers of Syrian Arabic and French. Syrian Arabic differs from English in having no phonologically overt exponent of indefiniteness; French differs from English in requiring phonologically overt exponents of definiteness/indefiniteness in all contexts. Evidence was collected from participants (including a control group of native speakers) through a forced-choice elicitation task, an oral story re-call task and a written production task. Results suggest that both Syrian Arabic and French speakers use English articles differently from speakers of LIs that lack articles, and differently from each other. Neither group shows evidence of fluctuating between definite and specific interpretations of articles (unlike speakers of article-less LIs), but the Syrian Arabic speakers in particular appear to have divergent knowledge of article distribution by comparison with the French speakers. It is argued that these findings are consistent with Full Transfer of the properties of the L1 initially, followed by restructuring towards target use of English articles, consistent with Full Access to Universal Grammar. Persistent non-target-like use of articles appears to be a problem of `feature reassembly'.

The research into aspects of EAP covers the twenty-five-year period from 1972 to 1997. The main focus of the submission is the use of needs analysis in relation to EAP, the implications of the findings, and their implementation with regard to teaching materials, course design and methods of teaching. Various methods are used to investigate the language and study skills needs of international students, mostly postgraduates at Manchester University: these include - language tests, self-assessment, questionnaires and surveys, observation and monitoring in class. Evaluation, feedback and follow-up investigations are used to assess the effectiveness of EAP courses and materials. The research covers the common core or study skills strand of EAP which includes the main study skills of listening and note-taking, academic writing, academic reading, speaking and discussion, reference skills, and examination skills. In addition, it covers subject-specific areas, exemplified here through Economics. One of the results of the investigations and the trial of teaching materials is the production of several books catering for students' needs in the main areas of study skills, especially academic writing. Co-operation with the Department of Economics enabled a framework to be developed which was used as the basis for designing and constructing materials for postgraduate students of Economics who needed to improve their study skills and English for Economics. The result of the co-operation was an improved language service for the students and the joint production of a book for their use. One of the main methods of collecting data for use in analysing the needs of students was the structured questionnaire and survey. The findings enabled the needs of students to be prioritised, and enabled EAP courses to be structured so that they met the students' changing needs. Course design was assisted by conducting the first national survey of EAP courses in British universities. Although the main focus of the research is upon the needs of students, the needs of EAP teachers are not overlooked. It had been discovered years ago that no single book catered specifically for the needs of EAP teachers. This had been -confirmed while writing a state-of-the-art article on EAP. Consequently, the first guide and resource book for EAP teachers and those undergoing training to become EAP teachers was written, and published in 1997.

Collocation learning has long been recognised as a problematic domain for even high-level learners and acknowledged to lag far behind acquisition of other SLA aspects. This thesis explores the role of vocabulary growth in the learning of L2 collocations. It addresses the relationship between vocabulary increase and L2 collocation learning, aiming to identify whether increasing lexemes in a syn(onym)set (Fellbaum, 1998) is the main factor responsible. A corpus-based cross-sectional study was undertaken on Chinese EFL learners’written production of three types of collocations: verb + noun collocations (the main research target), adjective + noun and noun + noun collocations. Lexical verbs in verb + noun collocations were classified into synsets and analyses were performed on collocations within these synsets. It finds a lag in L2 learners’verb + noun collocational knowledge with rising proficiency. This lag in collocational knowledge was found to be associated with an increase in lexical verbs learnt at higher levels. Collocation errors were seldom made where there was no increase in verb synsets. However, for synsets in which there was a verb increase, collocation errors involving new verbs were significantly more likely than errors with old verbs. The occurrence of collocation errors became increasingly limited to synsets with a verb increase as learners proceeded to more advanced levels. An alternative explanation was attempted to see if newly acquired nouns were also a factor responsible for the collocation lag. Results showed that in the majority of new nouns produced by higher levels of learners, collocations were target-like, and the percentages of new nouns in erroneous verb + noun collocations remained constant at both higher levels. In contrast, it finds an improvement in L2 learners’ knowledge of adjective + noun and noun + noun collocations. The thesis attempted to account for such differing performance from the perspective of vocabulary growth within synsets. The decreasing synonym density of verbs, adjectives and nouns may account for the relatively poorer performance on verb + noun collocations, and better performance on adjective + noun and noun + noun collocations. These findings are discussed with a view to a clearer understanding of the process of second language collocation learning, and to pedagogical implications.

Chinese allows object drop in contexts where there is an antecedent (anaphoric contexts), where English generally requires an overt object pronoun (e.g. Mary’s bike is broken. I am going to repair *(it) for her). In non-anaphoric contexts, however, English allows a null cognate object e as in Mary reads [e] every night whereas Chinese requires an overt cognate object (kan-shu, literally ‘read-book’). Previous SLA studies indicate Chinese learners of L2 English have problems unlearning anaphoric object drop in English, generally ascribed to effects of L1 transfer. This study brings a novel perspective to the L2 learnability problem by incorporating Cheng and Sybesma’s (1998) proposed negative correlation between the two rules: that allowing object drop in anaphoric contexts is incompatible with allowing object drop in non-anaphoric contexts.

In this dissertation I explore the nature of English as a Second Language (ESL) trainee teacher talk and the challenges trainees face in understanding and implementing it. The study takes a sociocultural perspective and uses a qualitative, multiple case study approach of nine non-native and native speaking trainees who are studying on a TESOL Master's programme which includes a microteaching and overseas teaching practice placement. The data set includes lesson transcripts which are analysed using Conversation Analysis and protocols of stimulated recall interviews which are analysed using the Threshold Concept framework. The data provide robust evidence of episodes of trainee talk and trainees' justification for use or non-use of interactional features. Findings show that trainees use talk to manage themselves rather than to manage pedagogy. They avoid interactional uncertainty, use unmarked mode switches, design turns which benefit the self and use rapport building features to confirm effectiveness. These features manifest in a closing down of communication. Findings also show that there is a hybrid nature to trainee identity in microteaching. Trainees shift between, for example, being a teacher and being a classmate and this leads to an opening up of interaction which is not evident in the teaching practice data. Furthermore findings show that trainees understand teacher talk in different ways: as a loss of familiar assumptions, a mimicking of 'correct' use and as a skill which needs to be performed in front of observers. The study has three main outcomes: a model showing trainee teacher talk features which 'open up' interaction and those that 'close down' interaction; a teacher talk concept 'map' which helps teacher educators explain the components which comprise teacher talk and a flow chart which bridges the gap between the theory of teacher talk as a threshold concept and the process of practising it.

Role play is commonly regarded as a useful activity for developing oral skills (Livingstone 1983; Maley 1987; Magos and Politi 2008) however, little research has been conducted on different types of role-play and on its use in EOP contexts, which is what the present study seeks to redress. This study reports on research investigating the effectiveness of two types of role-play, SSRP (Semi-scripted Role-play) and SSRP-NSRP (Semi-scripted Role-play followed by Nonscripted Role-play), in the teaching of Tourism English at a university in Korea. The study used an action research approach (Kemmis and McTaggart 1988) with the aim to improve Tourism English learners' oral skills.- Second -year students from different years participated in-Studies -I and II. In Study I students practised SSRP, while in Study II they engaged in SSRP-NSRP.

This research investigated the effectiveness of group work (GW) in EFL vocabulary learning by second . year, non-English major, university students in Taiwan, in comparison with working individually (IW). It also sought to find out students' attitudes towards vocabulary learning in each mode. The context was vocabulary-focused tasks following reading comprehension. In order to gather in-depth information about the process of, vocabulary discovery and retention and students' attitude to learning vocabulary through these two modes, questionnaires (44 students) and interviews (24 students) were used, along with vocabulary knowledge tests to measure knowledge gains (44 students). The design was repeated measures. Findings showed that students' overall improvement in vocabulary knowledge with GW was higher than that with IW. The results inferred from participants' statements in interviews and choices on the questionnaires showed that students' discovery of word knowledge in GW was greater than that in IW mainly because of having more opportunities to give and receive word knowledge from group members. Their retention of word knowledge in GW was a little higher than in IW because of giving and receiving helps during group discussion and learning in a friendly environment. Moreover, working with different group members in each GW. activity . seemed to affect their construction of groups (e.g. communicating about the tasks, assigning roles). Finally, students' attitude towards learning vocabulary through GW and IW changed. They favoured learning vocabulary through GW rather than IW less after GW was implemented, even though their vocabulary learning was improved. Pedagogical implications drawn from our findings were twofold for the educational authorities, English department, and teachers in Taiwan. First, the use of group work needed be encouraged in English class in secondary school and university. Second, provision of a vocabulary learning course at university and emphasis vocabulary learning in the university reading class were needed.

This study sought to describe and evaluate the 'process' of instructed English vocabulary learning in Tabika Perpaduan CTP) Pre-Schools for Malaysian children from low income families with regards to words presented to the pre-school children, teachers' vocabulary instructional methods and students' vocabulary achievement. Data collection was carried out by means of questionnaire for teachers, classroom observations of four teachers' instructional vocabulary methods at four pre-schools, and a series of one-to-one vocabulary tests for twenty-two students in these four pre-schools. Findings of the study have revealed, firstly, that the main sources of word learning were the two textbooks used by all TP pre-schools, and a variety of other commercial and teacher adapted language materials. It was also noted that many nouns were presented, with some verbs, adjectives and very few adverbs, and that many words were introduced to the students but with inadequate repetitions. Secondly, teachers used a variety of teaching techniques to present word meaning and to teach · listening vocabulary, reading vocabulary, speaking vocabulary and writing vocabulary. It is also found that teachers' instructional practices did appear to affect the students' achievement in the vocabulary tests. Thirdly, the students' mean vocabulary achievement revealed that the children had a greater ability to hear and read words, than to say and write words (with writing vocabulary being the weakest of their vocabulary knowledge), and mean vocabulary achievement differed across lexical fields. Analysis of students' spoken and written production mistakes revealed the students' developmental stages of language learning, the influence of Ll, and the different abilities in being able to spell words. The students' vocabulary achievement also showed that TP pre-schools were lacking in preparing the students to meet the guidelines/requirements stipulated by the National Pre-school Standard Curriculum, and the requirements for English vocabulary learning for Primary Year One.

This study explored the relationship of complex working memory (WM) and phonological short-term memory (PSTM) to aspects of second language (L2) oral production and self-repair behaviour. The study drew on Levelt's (1989; 1983) model of speech and perceptual loop theory of monitoring while the concept of WM was based on Baddeley and Hitch's (1974) multicomponent model of WM. Complex WM refers to the cognitive capacity of simultaneous storage and processing of information while PSTM refers to the capacity of the phonological store. The participants were 84 Emirati female university students learning English in an intensive language program in Abu Dhabi. It was hypothesised that given the limited automaticity of the language building processes in less advanced EFL learners, and thus the dependence of these processes on attentional resources, speakers with higher WM and PSTM scores would perform better in terms of fluency, accuracy, lexical and syntactic complexity in a task with simultaneous online planning. In addition, a relationship of WM with the number and the types of overt self-repairs was anticipated based on the attentional demands of the monitoring processes. Complex WM was measured with a backward digit span test in patticipants' Ll and a listening span test in L2. Phonological STM was measured with a simple word-recall test in L2. Statistical analysis of the data showed a relationship of complex WM with disfluency and general grammatical accuracy, while PSTM correlated significantly with speech rate, general and specific measures of grammatical accuracy as well as lexical variety. Complex WM and PSTM were also found to correlate moderately with overall oral performance scores. No statistically significant results emerged between complex WM, PSTM and number of self-repairs, but there was a significant negative correlation between PSTM and phonological error-repairs. Overall, the findings support that WM contributes to variation in L2 oral production but not overt self-repair behaviour.

Starting from the work of Levin and Rappaport Hovav (1995, 2005), this thesis gives an account of the lexical and morphological properties that are relevant to the syntactic expression of causative and inchoative events in Arabic and English, and reports a bidirectional study of how second language (L2) learners of these two languages acquire the different properties involved. It is argued that the two languages differ in both types of property. Given this, by comparing L1 Arabic-speaking learners of L2 English with L1 English-speaking learners of L2 Arabic, it is possible to address questions about the effects of L1 influence and the potential role of innate linguistic knowledge (Universal Grammar) on the development of L2 learners' 'interlanguage' grammars. The Full TransferIFull Access hypothesis of Schwartz and Sprouse (1994, 1996) predicts that the entirety of the L1 lexicon transfers at the initial state. At stages beyond the initial state, L2lexical properties will be acquired, yet L1 influences will persist. Results from grammaticality judgment and forced-choice elicitation tasks conducted with L2 speakers, involving a range of verb types with the potential to express causative/inchoative event meanings, show that low-proficiency learners who are no longer at the initial state nevertheless transfer lexical properties from their L1 into their 'interlanguage' grammars. Results from more proficient L2 learners at later transitional stages show that the two language groups follow different developmental patterns. Arabic learners of English produce overcausativisation and overpassivisation errors before they start to retreat from this overgeneralisation. In contrast, the English-speaking learners of Arabic are more target-like in their use of causatives/inchoatives at all proficiency levels. The implications of these findings for understanding the relative roles that the L1 properties, L2 properties and Universal Grammar play in the acquisition of lexical properties by L2 learners are discussed.