Session 4B

Recent Generative SLA (GenSLA) research has tried to account for the assembly of syntactic and semantic features of L2-English articles (Cho & Slabakova, 2014) and nouns (Choi, Ionin, & Zhu, 2018), but has done little to turn these results into practical teaching pedagogy (Whong, Gil, & Marsden, 2013a). This study seeks to fill this gap by creating a new instructional context informed by the findings and theories within GenSLA, termed linguistically-informed instruction, where teaching of L2-English articles and nouns is done through semantic universals: [±definite] for articles and [±atomic] and [±count] for nouns. Cho and Slabakova (2014) attribute the learning difficulty of [±definite] for English articles to the differences in non-morphological versus morphological expression of features in the L1 and L2. Choi and Ionin (2017) suggest that the difficulty in learning English nouns is influenced by [±atomic], as it is a semantic universal that is expressed differently in languages. Therefore, the present study seeks to follow-up on recent research to examine the learning task that L2-Mandarin, L2-English learners face when acquiring L2-English articles and noun types in three different learning contexts across in a pre- and post-test design. Three groups are included in the study: LING (n = 30) which uses semantic universals to teacher L2 articles and nouns, TRAD (n = 18) which receives instruction using their grammar book, and NOEX (n = 17) which receives no additional instruction. To evaluate the effect of instruction, participants were tested pre- (T0) and post-intervention (T1, T2) using an elicited-sentence imitation task and a forced-choice elicitation task. A linear mixed effects regression model was run investigating the interaction of type of instruction and time on the accurate production of article and noun type combinations. The model found the LING did not make significant improvements from T0 to T1 nor T2, but it did find that its gains were significantly greater than the other two groups from T0 to T1 (p < 0.05 in comparison to each both groups) and only in comparison to NOEX for to T2 (p < 0.05). With regard to the elicited-sentence imitation task, a linear mixed effects regression model was run investigating the interaction of type of instruction and time on the accurate correction of ungrammatical stimuli with regard to L2-English articles and noun types. The model found the LING group made significant improvement from T0 to T1 but not T2. When compared to other instructional contexts, its gains were significantly greater from T0 to T1 and T2 (p < 0.001 for all comparisons). Bearing in mind the significance of these results, if linguistically-informed instruction were implemented in a systematic way throughout a course, it may lead to greater gains when teaching complex linguistic concepts. Cho, J., & Slabakova, R. (2014). Interpreting definiteness in a second language without articles: The case of L2 Russian. Second Language Research, 30(2), 159-190. Choi, S. H., & Ionin, T. (2017). Acquisition and Processing of Mass Nouns in L2-English by L2 Learners from Generalized Classifier Langauges: Evidence for the Role of Atomicity. In M. LeMendola & J. Scott (Eds.), Proceedings of the 41st annual Boston University Conference on Langauge Development (BUCLD 2016) (Vol. 1, pp. 154-167). Somerville: Cascadilla Press. Choi, S. H., Ionin, T., & Zhu, Y. (2018). L1 Korean and L1 Mandarin L2 English learners' acquisition of the count/mass distinction in English. Second Language Research, 34(2), 147-177. Whong, M., Gil, K.-H., & Marsden, H. (2013a). Introduction: Generative Second Language Acquisition and Language Pedagogy. In M. Whong, K.-H. Gil, & H. Marsden (Eds.), Universal Grammar and the Second Language Classroom (pp. 1-13). Amsterdam: Springer Netherlands.

As classrooms are becoming increasingly diverse with the influx of work migrants, refugees, and other newcomers, English as Additional Language (EAL) teachers find themselves working in increasingly diverse educational settings. In fact, according to the European Commission (2013), teachers in Europe perceive working with multilingual students as one of the most urgent areas for professional development (PD). Consequently, there is a clear need to provide EAL teachers, supervisors, and program administrators working in multilingual contexts with better guidelines for lesson planning, delivery, and assessment. New approaches to language teaching must address and challenge the monolingual bias that continues to dominate many language classrooms (Hall & Cook, 2012), engage learners' whole linguistic repertoires (Cenoz & Gorter, 2014), enact "the multilingual turn" (May, 2014), and implement linguistically and culturally responsive teaching (Ladson-Billings, 1995; Lucas & Villegas, 2011) by promoting multilingualism as a core resource. Some instructional and assessment models already exist, such as the Sheltered Instruction Observation Protocol (Echevarría, Vogt, & Short, 2017), which emerged in the United States in the context of structured English immersion in K-12 and was designed to support English development. In this paper, we examine the effectiveness of PD offered to teachers at a multilingual school that serves newly-arrived immigrant and refugee students in Norway using the Multilingual Approach to Diversity in Education (MADE) observation tool. The PD and MADE were designed to foster multilingualism and support teachers, supervisors, and administrators working in linguistically diverse EAL contexts. MADE includes the following eight components: (1) Classroom as a multilingual space, (2) Interaction and grouping configurations, (3) Language and culture attitudes, (4) Language use: Learner, (5) Language use: Teacher, (6) Metacognition and metalinguistic awareness, (7) Multiliteracy, and (8) Teaching materials. With each component there is a subset of 2-7 indicators scored on a three-point Likert scale: 0 (not observed), 1 (observed), and 2 (observed multiple times). In this paper, we present data from fourteen observations of two teachers, which were conducted during two phases of a project that aimed to help the teachers develop adequate pedagogies for linguistically and culturally diverse classrooms. The following research question is addressed: To what extent does participation in PD change language teachers' pedagogical practices in multilingual EAL classrooms? The findings suggest that individual differences, such as family background, education, and teaching experience permeate the impact of PD. We discuss these results in the light of the PD in which the teachers participated and suggest ways in which the MADE can be incorporated into other contexts with culturally and linguistically diverse groups of learners as an assessment and PD tool.
References 
Cenoz, J., & Gorter, D. (2014). Focus on multilingualism as an approach in educational contexts. In A. Creese & A. Blackledge (Eds.), Heteroglossia as practice and pedagogy (pp. 239-254). Berlin: Springer.
Echevarría, J., Vogt, M., & Short, D. J. (2017). Making Content Comprehensible for English Learners: The SIOP Model, 5th Edition. Boston: Pearson. European Commission (2013). The teaching and learning international survey (TALIS). Main findings from the survey and implications for education training and policies in Europe. 
Hall, G., & Cook, G. (2012). Own-language use in language teaching and learning. Language Teaching, 45(3), 271-308. 
Ladson-Billings, G. (1995). Toward a theory of culturally relevant pedagogy. American Educational Research Journal, 32(3), 465-491. 
Lucas, T., & Villegas, A. M. (2011). A framework for preparing linguistically responsive teachers. In T. Lucas (Ed.), Teacher preparation for linguistically diverse classrooms (pp. 35-52). New York, NY: Routledge. 
May, S. (Ed.) (2014). The multilingual turn. Implications for SLA, TESOL, and Bilingual Education. New York: Routledge.

This paper shows the results of a classroom study aiming at exploring the instructional effects of structure-based (SB) or dynamic usage-based (DUB) instruction with free response, communicative writing tasks after three years of L2-French instruction on linguistic complexity measures in (morpho)syntax and lexicon. The fundamental difference between an SB and a DUB approach is the way language itself is viewed. An SB approach assumes language is a complex system in which different autonomous sub-components (such as syntax and lexicon) interact predictably according to 'rules'. In contrast, a DUB approach assumes language is a complex dynamic system in which there is no fundamental difference between syntax and lexicon, and that language is used on the basis of the speaker's individual routines. The term 'dynamic usage based' (DUB) is inspired by the title of one of Langacker's articles (2000), in which he argues that a usage-based view is per definition a complex dynamic systems theory view. These different views have implications for how language should be presented and instructed, and the behaviour that learners should aim for (Verspoor 2017). We investigated data from forty-three young high school beginner learners of L2-French after three years of instruction with similar amounts of L2 exposure. The SB treatment included a traditional focus on explicit grammar; the DUB group was taught using the Accelerated Integrated Method, a highly communicative, meaningfocused method without explicit instruction, but with a great deal of exposure and repetition to induce frequency effects. The effectiveness of two L2 teaching methods – SB and DUB – were evaluated using several (morpho)syntactic complexity measures such as sentence length and morphological complexity, and lexical measures such as diversity (Guiraud), word complexity (average word length) and multi-word (MW) sequences. Our hypothesis was that the SB group to outperform the DUB group on (morpho)syntactic linguistic complexity measures and the DUB group to outperform the SB group on MW sequences. Results after three years show that DUB instruction leads to more linguistic complexity in terms of various (morpho)syntactic and some lexical measures (multi-word sequences coverage). On other lexical measures (such as Guiraud index and average word length), no differences were found. The results will be discussed using insights from the dynamic usage-based perspective. References : Langacker, R. W. (2000) A dynamic usage-based model. In M. Barlow and S. Kemmer (eds) Usage-based Models of Language 1–63. Stanford: CSLI. Verspoor, M. (2017) Complex dynamic systems theory and L2 pedagogy. In L. Ortega and Z. Han (eds) Complexity Theory and Language Development: In Celebration of Diane Larsen-Freeman 143–62. Amsterdam: John Benjamins Publishing Company. https://doi.org/10.1075/lllt.48.08ver

Proficiency is one of the most important constructs in Second Language Research. Yet, its measurement has not always received the attention it deserves, and practices of proficiency level assignment have been the subject of continued criticism (e.g. Hulstijn et al. 2010). In Learner Corpus Research, more particularly, Carlsen (2012) noted that learner corpus compilation projects still relied on variables such as ‘institutional status’ or ‘year of study’ to assign a proficiency level to learners (and even learner groups) while these external or learner-centred criteria are largely regarded as unreliable (Thomas, 1994). To address these criticisms, the AndreSpråksKorpus (ASK) compilers were among the first to adopt a systematic a posteriori text-based approach to proficiency grounded in the professional field of language assessment. The ASK texts were rated by a minimum of five trained raters; estimates of rater severity and rater reliability were also used to provide feedback to the raters and evaluate the rating procedure (see Carlsen, 2012 for more details). While this approach is undoubtedly exemplary, the drawback is “the time and costs required” (ibid.:179). The main objective of the Crowdsourcing Language Assessment Project (CLAP) is to investigate whether crowdsourcing can offer practical solutions to the time and cost difficulties associated with a text-based approach to proficiency assessment in learner corpus research. Crowdsourcing refers to the practice of soliciting contributions from a large group of people or the general public rather than from traditional employees or suppliers. In this specific project, we are soliciting evaluations of argumentative essays written by English as a Foreign Language (EFL) learners from members of the European Network of Combining Language Learning with Crowdsourcing Techniques (enetcollect; Lyding et al., 2018) instead of professional raters. The project relies on the technique of adaptive comparative judgement (ACJ; Pollitt, 2012), which is a modification of Thurstone’s method of comparative judgement that has been reported to produce reliable and valid assessment of written performances (Van Daal et al., 2019): learner texts are presented in pair to raters who have to choose the ‘best’ one. This technique is based on the assumption that people are able to compare two performances more easily and reliably than to assign a score to an individual performance (Lesterhuis et al., 2017). A second critical assumption underpinning CJ is its reliance on holistic judgment: “Judges do not receive criteria to guide their judgment process, but only a general description regarding the writing competence to be assessed” (Van Daal et al, 2019). By means of an iterative and adaptive algorithm, the ACJ platform used (here ComPAIR, Potter et al., 2017) then produces a scaled distribution of student performances. The study design was piloted locally in 2019 and we will launch the project in February 2020. Unlike in traditional crowdsourcing tasks, we will ask raters to fill in a short questionnaire about their familiarity with English and the task of writing assessment. The results will make it possible to answer the following research questions: - Can a crowd of people be used to assess learner texts reliably and validly? - How do the scores generated by the crowdsourced language assessment task compare with the scores assigned by experts to the same learner texts? - What is the impact of different language skills and language assessment expertise on the task of language assessment? We believe addressing these questions in the fields of SLA and LCR is of crucial importance given the centrality of the construct of proficiency. If results are satisfactory, it is our hope that the approach be used by the SLA/LCR crowd to add text-based proficiency levels to a variety of learner corpora for different languages.