Corpus linguistics

The Applied Linguistics program at NAU is one of the preeminent programs in corpus linguistics in the world. For more than 25 years faculty and students have produced groundbreaking research in corpus linguistics. This trend continues as evidenced by the research currently underway at NAU.

Doug Biber and Randi Reppen recently obtained a $1.6 million US Institute of Education Sciences grant with Trina Spencer (University of South Florida) to develop a corpus of language elicited from children in grades K-3. Their goal is to examine the children's vocabulary development and how children who are below, at, and above grade level vary in their performance on various spoken and written language tasks.

Jesse Egbert and Doug Biber are currently collaborating on three major research projects. First, using their Corpus of Online Registers of English (CORE; Biber et al., Reference Biber, Egbert and Davies2015; Egbert et al., Reference Egbert, Biber and Davies2015; Biber & Egbert, Reference Biber and Egbert2018), they are undertaking a comprehensive description of online registers (Biber & Egbert, Reference Biber and Egbert2016; Egbert & Biber, Reference Egbert and Biber2016). This has led to a reformulation of the construct of register in continuous rather than categorical terms (see Biber et al., Reference Biber, Egbert and Kellerforthcoming). Second, using the new Spoken British National Corpus 2014 (Love et al., Reference Love, Dembry, Hardie, Brezina and McEnery2017), Egbert and Biber, along with doctoral candidates, Daniel Keller and Stacey Wizner, are collaborating with researchers at Lancaster University to develop methods to segment conversations into smaller analyzable units based on communicative functions in order to more precisely explore linguistic variation in sub-registers of conversations. Third, Egbert and Biber are directing the development of a new part-of-speech tagger, developed principally by doctoral candidate Brett Hashimoto and recent MA-TESL alumnus Michael Suhan. This new tagger is built on top of the CLAWS tagger and improves on Biber's original tagger with new grammatical tags, improved systematicity, increased processing speed, and an online interface.

Bill Crawford is using his Corpus of Collaborative Oral Tasks (CCOT), a corpus of 775 spoken tasks carried out by dyads of L2 learners of English, to better understand how task parameters affect language production in oral collaboration tasks. He is also currently editing a book titled Multiple perspectives on learner interaction: The corpus of collaborative oral tasks. The book draws on the CCOT and includes a number of studies that are relevant to SLA and assessment (Crawford, Reference Crawfordforthcoming). Finally, Brett Hashimoto and Kyra Nelson (Reference Hashimoto and Nelson2020) have applied the corpus linguistics approach to pursue validity evidence of a common research instrument in L2 pragmatics called the discourse completion test (DCT). They analyzed features of advice-giving expressions using the Q+A corpus (Baker & Egbert, Reference Baker and Egbert2016) and compared the features against those elicited from 33 participants using a DCT.

Luke Plonsky has recently developed the Second-language Research Corpus (L2RC), a corpus of approximately 22,000 articles spanning several decades from 22 prominent L2 journals. He plans to use this corpus to examine linguistic as well as methodological and theoretical trends in the field.

Much of the corpus linguistic research at NAU triangulates corpus methods with other linguistic methods. Jesse Egbert has recently completed two co-edited books on this topic (Baker & Egbert, Reference Baker and Egbert2016; Egbert & Baker, Reference Egbert and Baker2019). Ph.D. student Daniel Keller is carrying out his dissertation project on psychological evidence of register by triangulating psycholinguistic and corpus methods. In addition, a recent seminar taught by Egbert spurred a wide range of novel applications of methodological triangulation and corpus linguistics by Ph.D. and M.A. students in the program (Keller & Kostromitina, Reference Keller and Kostromitina2020).

L2 speech

Under the direction of Okim Kang, the Applied Linguistics Speech Lab has been engaged in the development and application of computer systems for prosody modeling and automatic fluency extraction. Based on Brazil's (Reference Brazil1997) intonation model, Kang's patented prosody modeling system detects phones, rather than words, and uses these phones as the foundation for identifying prominent syllables, tone units, and tonic syllable pitch (see Kang & Johnson, Reference Kang and Johnson2018a, Reference Kang and Johnson2018b).

Within language teaching and learning, these automatic speech recognition systems are useful for speech assessment and linguistic analysis. For instance, Kang and Kevin Hirschi (with Helmer Strik and Catia Cucchiarini at Radboud University, John Hansen at the University of Texas-Dallas, and Keelan Evanini at the Educational Testing Service) are currently developing an affective interactive oral communication tutor for limited English proficiency populations. Using the 2019 IELTS joint research fund, Kang (with Seoul National University's Hyunkee Ahn and NAU's Meghan Moran and Katherine Yaw) is also applying such tools to a project analyzing the linguistic progression of Korean adult learners of English who are preparing for the IELTS (Kang et al., Reference Kang, Thomson and Moran2018a, Reference Kang, Thomson and Moran2018b, Reference Kang, Thomson and Moran2019; Kang et al., Reference Kang, Ahn, Moran and Parkforthcoming). In an additional study involving speech analysis, alumna María Nelly Gutiérrez Arvizu (Universidad de Sonora) is investigating fluency features in Spanish-speaking kindergarten learners of English. This is the first funded project supported by the program of Alianza Inter-Universitaria Sonora-Arizona.

In terms of listener perception of speech, recent projects have focused on acceptability judgments of speech varieties in different professional and educational contexts. Okim Kang and her research team have looked at the relationship between speaker intelligibility and perceived acceptability for different professions, as well as English language teachers' perceptions of the acceptability of World English varieties as models in the classroom (Dalman & Kang, Reference Dalman and Kang2019). In addition, Mohammedreza Dalman recently examined undergraduate listeners' perceptions of TOEFL iBT speaking performances, finding that high-scoring speakers on the TOEFL iBT were judged at varying levels of acceptability for teaching in a university context. Wheeler and Kang are currently investigating the speech of L1 Spanish/L2 English speakers as perceived by recent returnees of study abroad programs to Spanish-speaking countries.

Finally, there are several projects focusing on interactive speech. In the speech lab, for example, Mark McAndrews is working on a fluency and speech alignment analysis of university-level learner speech using Crawford's (CCOT). Kang is working with Naoko Taguchi to analyze intonation features for pragmatic function (Kang, Kermad, & Taguchi, Reference Kang, Kermad and Taguchiunder review). Garrett Larson's dissertation study has been developing measures for assessing collaboration in speaking assessments and examining the impact of interactional features on speaker performance. Though not part of the speech lab, Soo Jung Youn and Shi Chen have been investigating interactional competence and rater cognition in paired speaking assessments (Youn, Reference Youn2018; Youn & Chen, Reference Youn and Chen2019). Currently Naoko Taguchi and Soo Jung Youn are writing a survey of speaking tasks used in L2 pragmatics research (Taguchi & Youn, Reference Taguchi, Youn, Derwing, Munro and Thompsonforthcoming).

Research methods

With a heavy focus on quantitative research methods, NAU arguably represents the hub of the methodological reform movement currently taking place in applied linguistics (see Gass et al., Reference Gass, Loewen and Plonskyforthcoming; Marsden & Plonsky, Reference Marsden, Plonsky, Gudmestad and Edmonds2018). We will highlight two major strands of NAU student and faculty contributions to this movement.

First, NAU students and faculty have sought to help the field better understand and improve quantitative data analyses. Building on earlier work from both outside and within the field challenging an overreliance on null hypothesis significance testing, two different approaches have been taken. One involves critically examining existing/familiar techniques such as measurement reliability (McKay & Plonsky, Reference McKay, Plonsky, Winke and Brunfautforthcoming), the use of grades as proxies for L2 proficiency or achievement (Brown et al., Reference Brown, Plonsky and Teimouri2018), the use and interpretation of statistics such ANOVA and multiple regression and their corresponding effect sizes (partial) eta² and R ² (Norouzian & Plonsky, Reference Norouzian, Plonsky, Phakiti, De Costa, Plonsky and Starfield2018a, Reference Norouzian and Plonsky2018b; Plonsky & Ghanbar, Reference Plonsky and Ghanbar2018), and the checking and reporting of statistical assumptions (Hu & Plonsky, Reference Hu and Plonskyforthcoming). Complementary to this work, we have been active in introducing and promoting the use of statistical techniques new to applied linguistics such as bootstrapping (Egbert & LaFlair, Reference Egbert, LaFlair, Phakiti, De Costa, Plonsky and Starfield2018; Egbert & Plonsky, Reference Egbert, Plonsky, Gries and Paquotforthcoming) and Bayesian data analysis (e.g., Norouzian et al., Reference Norouzian, de Miranda and Plonsky2018, Reference Norouzian, de Miranda and Plonsky2019).

A second main emphasis of the methodological work at NAU has involved research synthesis and meta-analysis. Recent and ongoing meta-analyses have addressed, for example, the effects of structural priming (in preparation), L2 pragmatics instruction (Plonsky & Zhuang, Reference Plonsky, Zhuang and Taguchi2019), L2 suprasegmental instruction (McAndrews, Reference McAndrews2019), the effects of L2 pronunciation instruction (Saito & Plonsky, Reference Saito and Plonsky2019), and the relationship between working memory and reading comprehension (Shin, Reference Shinunder review). However, two other types of efforts are also part of this strand. Students and faculty in the program have produced a number of methodological syntheses, which apply meta-analytic techniques as a means to describe and evaluate research and reporting practices related to a given domain or technique such as self-paced reading tasks (Marsden et al., Reference Marsden, Thompson and Plonsky2018), grammaticality judgment tasks (Plonsky et al., Reference Plonsky, Marsden, Crowther, Gass and Spinnerforthcoming), and outlier handling techniques (Nicklin & Plonsky, Reference Nicklin and Plonskyforthcoming). In addition to teaching semester-long courses as well as shorter workshops and bootcamps, Plonsky has also sought to improve and advance the field's use of meta-analytic procedures. One current project in this area involves applying corpus linguistic techniques to research synthesis as means to automatically extract, for example, demographic features or a particular effect size across a sample of studies.

SLA

There is a strong and long-standing tradition of SLA research at NAU. Currently, several projects focus on the role of individual differences (IDs). In addition to recent works on established IDs such as learning strategies (Plonsky, Reference Plonsky, Chamot and Harris2019; Plonsky & Sudina, Reference Plonsky, Sudina, Mohebbi and Coombeforthcoming) and anxiety (Teimouri et al., Reference Teimouri, Goetze and Plonsky2019), a series of studies is underway on the role of ‘grit’, defined as perseverance and passion for reaching long-term goals (Duckworth et al., Reference Duckworth, Peterson, Matthews and Kelly2007). Specifically, Teimouri et al., (Reference Teimouri, Plonsky and Tabandehunder review) have recently developed and proposed an L2-Grit Scale—the first language-specific scale designed to measure grit in L2 learners—and found it to be positively related to L2 achievement (rs = .27–.35). Subsequent and ongoing projects further explore (a) the role of grit among L2/L3 (third language) learners in Russia (Sudina & Plonsky, Reference Sudina and Plonskyunder review), (b) the relative contributions of L2 grit and aptitude in predicting achievement (Tabandeh et al., Reference Tabandeh, Teimouri and Plonskyunder review), and (c) a new measure of occupation-specific grit—the L2-Teacher Grit Scale—that demonstrated stronger associations with retention and years of English as a second language/English as a foreign language (EFL) teaching relative to Duckworth and Quinn's (Reference Duckworth and Quinn2009) domain-general grit scale (Sudina et al., Reference Sudina, Vernon, Foster, del Villano, Hernandez, Beck and Plonskyunder review). Adding to the research on understudied ID variables, Naoko Taguchi has investigated the impact of intercultural competence on L2 learners' pragmatic competence (Taguchi, Xiao, & Li, Reference Taguchi, Xiao and Li2016).

SLA research underway at NAU also addresses a number of other topics. For example, Zhang and Crawford (Reference Zhang and Crawfordunder review) have been investigating the use and resolution of language-related episodes (LREs) in collaborative writing. The findings indicate that the L1 interaction condition results in more LREs and more correctly resolved LREs. In addition, Larson (Reference Larsonunder review), taking a task-based approach, is examining the speech transcripts produced by learners in the context of pre-task planning to identify interactional behaviors (e.g., functions of talk, interaction features) and their associations with speaking scores. Taguchi has been examining task condition effects on L2 speech acts by analyzing L2 English learners' speech acts elicited through two different speaking tasks: virtual reality-based and computer-based. Furthermore, Lee et al., (Reference Lee, Plonsky and Saitoforthcoming) have recently tested the relative effects of perception- and production-based pronunciation instruction among learners of EFL.

Psycholinguistics

Research in progress at the Language and Memory Lab at NAU is building on research in L2 morphology learning. Vedran Dronjic, Daniel Keller, Masha Kostromitina, and Maren Greve are conducting a large-scale investigation of the interaction between IDs in learning and memory abilities, morphological type, and input variability on the learning of morphological systems that differ from English. Learners in the high-variability condition receive input containing a wider range of nouns occurring with the target set of morphemes than learners in the low-variability condition. This project promises to illuminate how IDs in memory and learning abilities interact with input when learning different types of morphology.

Morphological transfer and input variability are also under investigation in a multi-site, NSF-funded neurolinguistic investigation of morphology learning. Vedran Dronjic and collaborators at Vanderbilt University and the University of Haifa are using chronometric and hemodynamic (fMRI) techniques to investigate the learning of concatenative and root-and-pattern derivational morphology by native English and Hebrew speakers. The researchers hypothesize that (1) Hebrew speakers will benefit from L1-L2 transfer for root-and-pattern morphology, evidenced by a faster transition from (para)hippocampal to neocortical representation, (2) morphologically complex words that are structurally similar in the L1 and L2 will exhibit fronto-striatal (compositional) representation, while those that are dissimilar will rely on temporal regions implicated in lexical representation, and (3) greater variability in the training stimulus set will result in better learning.

IDs in working memory have also been the focus of a pair of studies at NAU in relation to L2 reading. Building on work described in Shin et al. (Reference Shin, Dronjic and Park2018) and using a newly developed computerized task that taps storage, processing accuracy, and processing speed with millisecond precision, Jihye Shin is exploring how the subcomponents of working memory facilitate or limit the application of background knowledge during L2 reading. Shin's work is the first to explore the effect of working memory differences on the application of background knowledge at this level of granularity.

Investigation into the effect of non-linguistic context on language processing is currently underway at NAU as well. Situational (i.e., register) variation is thought to drive linguistic variation (Biber & Conrad, Reference Biber and Conrad2019), but little is known about how situation might impact language processing. Daniel Keller and Vedran Dronjic are using a new type of primed lexical decision task to investigate the effect of register expectations on lexical access speed. Preliminary results (Keller & Dronjic, Reference Keller and Dronjicunder review) suggest that participants are quicker to recognize register keywords when they have been primed by a visual scene associated with the register. Ongoing work expands on these findings by testing the effect with registers with different lexical features and treating keyness as a linear predictor of response time. This research promises to expand our understanding of how language users produce situationally appropriate language.