Session 2B

A key problem for second language (L2) learners in many contexts is to improve the fluency of their speech (Tavakoli & Wright, 2020). While fluency practice should be an integral part of the curriculum (Nation, 2007), it is often not included in L2 curricula. Several researchers (e.g. Goncharov, 2019; Liang, 2017; Martinez & Schmitt, 2012) have highlighted the benefits of introducing fixed expressions – i.e. formulaic sequences (FSs) – for improving fluency and called for their inclusion in L2 curricula. However, there are still very few in-depth studies that examine the complexity, accuracy and fluency (CAF) of the FSs found in L2 speech in detail. The current study fills an important gap in our understanding of bigrams and trigrams in spontaneous L2 speech in that we developed a novel complexity index for FSs on the basis of indices from TAALES (Kyle et al., 2018), and monolinguals' judgements of the transparency of each FS and their complexity. The target population of this study are adult L2 learners of English in Kuwait. The sample under study (N=51) are all at A1 to A2 level according to the CEFR. The specific focus of the current paper is on the computation of CAF of FSs. We use our novel index of complexity to study the complexity of the FSs, and evaluate the fluency of FSs by studying pauses before and within each FSs. In addition, we focus on errors and error-free chunks in FSs, subsequently analysing the CAF of L2 learners' overall speech samples, and correlating this with the CAF of the FSs. Results from the pilot study show that CAF of students' speech correlate significantly with participants' use of FSs in a speaking task, revealing a clear link between overall fluency of L2 learners' speech and the CAF of FSs. 



References
Goncharov, G. (2019). The Effect of Direct Instruction in Formulaic Sequences on IELTS Students' Speaking Performance. Advanced Education, 6(11), 30-39.

Kyle, K., Crossley, S., & Berger, C. (2018). The tool for the automatic analysis of lexical sophistication (TAALES): version 2.0. Behavior research methods, 50(3), 1030-1046.

Liang, H. (2017). An Empirical Study on the Effects of Computer-Corpus-based Formulaic Sequences on College Students' Oral English Learning. International Journal of Emerging Technologies in Learning (iJET), 12(08), 67-76.

Martinez, R., & Schmitt, N. (2012). A phrasal expressions list. Applied linguistics, 33(3), 299-320.

Nation, P. (2007). The four strands. International Journal of Innovation in Language Learning and Teaching, 1(1), 2-13. 

Tavakoli, P., & Wright, C. (2020). Second language speech fluency: from research to practice. Cambridge University Press.

Fluency is a key construct in communicative speaking ability, reflecting automaticity and speed of the necessary speech production processes. Consequently, aspects of fluency are apparent in most speaking assessment criteria. Research in the field of second language acquisition has shown that a number of fluency aspects are related to judgements on fluency (Suzuki et al., 2021). These include aspects of speed of speech, pausing, and repair.
For research into second language acquisition as well as for research linking proficiency levels to aspects of fluency, raters are currently usually told to base their judgments on specific pre-defined utterance fluency characteristics. In this study, we examine whether there are other speech factors that fall under the concept of fluency. 
In this (mainly) qualitative study, 23 native speakers of Dutch took part in a rating session, rating samples of speech by Dutch L2 speakers, taken from Bosker et al. (2013). Two groups of raters were juxtaposed: one group received a definition of (narrow) fluency only, which mentioned "smooth translation from thoughts to speech" (intuitive instructions group); the other group had to judge fluency based on the same definition, which was now followed by mentioning specific fluency characteristics (fluency instructions group, as in Bosker et al., 2014). To investigate whether intuitive ratings are based on more or other aspects than the instructed ratings, 15 participants filled out an on-line questionnaire, and 8 participants carried out a stimulated recall task after they finished the ratings. In these questionnaires and stimulated recalls, the raters were asked to reflect on how they came to their judgements on fluency (cf. Brown et al., 2005). 
The questionnaires and transcripts of the stimulated recalls were analyzed by combining deductive and inductive methods of categorization. In total, 17 different fluency categories emerged from these analyses. Results indicated that the intuitive instructions group was less likely to have disfluencies influence their ratings negatively, but was more inclined to judge overall proficiency rather than fluency in the narrow sense. Additionally, both groups showed a sensitivity for pause distribution that helped them to determine the speaker's ease of lexical retrieval. Both groups mentioned aspects of intonation and planning efficiency as essential components of fluency. We conclude our presentation with implications for fluency rating instructions for language assessment and second language acquisition research purposes.


Bosker, H. R., Pinget, A. F., Quené, H., Sanders, T., & De Jong, N. H. (2013). What makes speech sound fluent? The contributions of pauses, speed and repairs. Language Testing, 30(2), 159-175.
Brown, A., Iwashita, N., & McNamara, T. (2005). An examination of rater orientations and test-taker performance on English for academic purposes speaking tasks (TOEFL Monograph 29). Princeton, NJ: Educational Testing Service.
Isaacs. T. & Thomson, R.I. (2013). Rater Experience, Rating Scale Length, and Judgments of L2 Pronunciation: Revisiting Research Conventions, Language Assessment Quarterly, 10 (2), 135-159, DOI: 10.1080/15434303.2013.769545.
Suzuki, S., Kormos, J., & Uchihara, T. (2021). The Relationship Between Utterance and Perceived Fluency: A Meta‐Analysis of Correlational Studies. The Modern Language Journal.

The role of silent pausing in speech production has received much attention in recent years. A key issue that has intrigued researchers has been whether speech processing underlying pauses may differ depending on pause location, as predicted based on speech production models (e.g., Levelt, 1989). To date, no direct evidence is available about the neural processes underlying silent pauses in L1 and L2 speech. While a growing number of studies have found evidence in support of this hypothesis, previous studies have solely used cognitive-behavioural tools to investigate pausing behaviours (e.g., De Jong, 2016). Therefore, this study aimed to examine the neural correlates of silent pausing, in particular, how location of pausing (mid- versus end-clause) may relate to neural processes during spontaneous speech production and how pausing-related neural processes may vary across L1 and L2 speakers. Based on previous behavioural research, we hypothesised that mid- and end-clause pauses would be linked, respectively, to activation in language- and conceptualisation-related brain areas during both L1 and L2 speech, with the difference being more pronounced for L2 speakers. 
The participants were 26 intermediate-level Japanese university students of L2 English, who carried out eight monologic oral decision-making tasks, four in L2 English and four in L1 Japanese. The tasks asked participants to resolve crisis-related problems (e.g., selecting items to take after a flood alert, choosing people to be vaccinated when limited vaccines are available). Language and task version were counterbalanced across participants. While the participants were completing the tasks, their brain activity was recorded through fMRI scanning. An MRI-compatible noise-cancelling microphone was used to record their performance. Participants' L1 and L2 speech was transcribed using the software PRAAT and was annotated for mid- and end-clause silent pauses. Brain imaging data, for the duration of silent pauses, were analysed by using SPM12 in the conventional two-level analysis. First, we modelled two regressors of mid- and end-clause pauses in each L1 and L2 language task for each participant as an independent variable. Second, for the group analysis, we tested the effect of pause location (mid- vs. end-clause) and language (L1 versus L2) with two-way ANOVA implemented in SPM12. 
Our preliminary analyses, based on 32 L2 speech samples from 8 speakers, indicate that, as predicted, the left interior frontal gyrus (language-related area) showed greater activation during mid-clause pauses, whereas the left precuneus and left inferior parietal lobule (conceptualisation-related areas) were more active when participants paused end-clause. We will discuss the implications of the study for models of L1 and L2 speech production and cognitive-neuroscience more generally. We will also discuss the benefits and challenges of triangulating cognitive-behavioural and neuroimaging data to assess pausing behaviours during speech production.


Levelt, W. J. M. (1989). Speaking: From intention to articulation. MIT Press.
De Jong, N. H. (2016). Predicting pauses in L1 and L2 speech: The effects of utterance boundaries and word frequency. International Review of Applied Linguistics in Language Teaching, 54, 113–132.

Linguistic performance elicited by language tasks has generally been operationalized in terms of complexity, accuracy and fluency (CAF). However, as has been argued in a number of studies (e.g., De Jong et al., 2012; Révész et al., 2016), assessment of L2 proficiency is impossible without considering the efficacy and appropriacy of L2 performance (henceforth 'functional adequacy', FA). From the perspective of task-based language assessment (TBLA; Long, 2015, 2016; Norris, 2016), FA is conceived of as a multi-layered, goal-directed, task-related construct, in terms of successful task completion by the speaker/writer in conveying a message to the listener/reader. A rating scale of FA for the assessment of oral and written performance has been developed, which distinguishes four dimensions: Task Requirements, Content, Comprehensibility, and Coherence & Cohesion (Kuiken & Vedder, 2017). 
In order to investigate the reliability, validity and applicability of the FA scale, a number of experimental studies have been conducted in which FA was assessed by both expert and non-expert raters, in different learning contexts, involving various source and target languages, proficiency levels (A2-C1), task types and modalities. Some of these studies have also investigated the relationship and mutual development of FA and CAF, resulting in mixed findings. The main outcome of the studies in which the FA scale was employed was that the FA scale is a reliable, valid and user-friendly tool and that, in terms of applicability, its scope is sufficiently broad. A number of issues and challenges for future research, however, still remain. 
The goal of our presentation is to discuss perspectives and challenges of research on FA for TBLA and SLA, in particular regarding the following topics: 
1. Standardization
An important issue concerns the reliability, validation and/or adaptation of the FA scale, in relation to learning context, target language, task type, task modality. In order to assure comparability of studies it is necessary to standardize test instrument, methodology, assesment tasks (use of 'proto-typical tasks'), data analysis and rater training.
2. FA in relation to (sub)components of CAF
Although (sub)components of FA and CAF appear to be connected to some degree, the overall picture is still unclear. Further investigation is needed, e.g., associations between FA descriptors and CAF measures, or the extent to which the relationship of FA and CAF is moderated by proficiency level and task type. 
3. FA in interactional tasks
So far, the FA rating scale has been employed exclusively for the assessment of monologic tasks. An important question is whether and how the rating scale can be used (adapted and/or extended) for interactional tasks. 
4. FA in classroom practice
Another issue which needs to be further explored is the role of FA in classroom and assessment practice, and how it can be incorporated into the field of instructed second language acquisition (ISLA). Future research should also examine the impact of different instructional treatments on the development of FA, and the possibility to use the scale for self-assessment by learners and/or peer feedback.