Theory-supported corpus pedagogy for ESL pre-service teachers: using Parallel EAP Corpora for language learning

2.1 Constructivist learning and socio-cultural theory

Constructivist learning has gained attention amongst educational researchers because of benefits of improved learning outcomes and the promotion of independent learning (Bada and Olusegun 2015). It considers learners as active agents in knowledge acquisition and requires them to engage in discovery learning to activate higher-order cognitive skills, such as inferencing and hypothesising. However, it may not be ideal for all learners because its underlying learning mechanism places considerable cognitive loads on learners (Kirschner et al. 2006; O’Keeffe 2021). Hence, integrating socio-cultural theory into constructivist learning approaches is recommended (Brown and Palincsar 1989; Kirschner et al. 2006; Palincsar 1998).

Vygotsky’s (1978) socio-cultural theory posits that knowledge is co-constructed through collaborative dialogues and social interactions with guidance and mediated support in the form of scaffolding. This theory, which emphasises the importance of providing extensive support and mediation during the learning process, has gained wide acceptance in educational technologies (Feyzi and Yasrebi 2020).

Constructivist learning and socio-cultural theory are also mainstream theoretical approaches for increasing CT’s effectiveness in language learning (O’Keeffe 2021). CT provides abundant, authentic language data and intensive exposure for learners to construct their second language (L2) knowledge independently (Flowerdew 2015; Lee et al. 2019), resulting in increased recall of word meanings and collocations (Lee et al. 2019), while teacher intervention helps to alleviate the unsatisfactory effects of individual exploration (Johansson 2009), and peer collaboration enhances learners’ academic performances and engagement (Huang 2011). However, there is a lack of pre-designed learning activities based on these theories (O’Keeffe 2021). The social constructivist learning theory (Brown and Palincsar 1989; Palincsar 1998), which acknowledges the overlaps and benefits of constructivist learning and socio-cultural theory, was adopted to support the CT-based pedagogical design in this study.

2.2 The noticing hypothesis and CT

Schmidt’s (1990, 2001 noticing hypothesis asserts that noticing precedes understanding and facilitates the conversion of input into intake. This hypothesis is closely intertwined with the frequency of input. As Ellis (2006) concludes, language processing is largely based on frequency and probabilistic knowledge. Schmidt (2010) also suggests that learners could notice the gaps in their language use by consciously comparing their output to input in the target language.

CT fosters learners’ awareness of target forms because inductive learning, a prominent approach in CT, relies heavily on noticing language features (Flowerdew 2015). CT provides ample data pertaining to target forms, including collocations, concordance lines, contexts and frequency information. By inferring language use patterns from authentic language data, learners can internalise statistical information about language use with the assistance of CT (Crosthwaite and Boulton in press). Specifically, Boulton (2011) found that corpus training facilitated learners’ development of noticing skills, and Gaskell and Cobb (2004) revealed that learners could apply noticing strategies to concordance lines for self-correction in writing.

However, existing CT studies have mainly focused on native speaker or expert writer corpora, with few having used learner corpora in classrooms (see Boulton 2021). Learner corpora, which consist of authentic language data obtained from ESL learners, give teachers and researchers valuable insights into the challenges that these learners encounter in academic writing (Granger 2004). Comparing expert and learner corpora can help students to overcome their linguistic challenges and can increase their motivation (Granger and Tribble 2014); their learning outcomes can be significantly improved (Ackerley 2017; Cotos 2014) by providing both expert and learner corpus data to enhance learners’ noticing of gaps in their language use (Schmidt 2010).

2.3 Applications of CT and reporting verbs

CT for academic writing and vocabulary learning has been widely introduced to learners. For example, Chen and Flowerdew (2018) focused on 37 studies in the academic writing context, and stated that CT was effective for learners (including Chinese students) in terms of error correction, vocabulary acquisition and learning the rhetorical features of academic writing. Lee et al. (2019) reviewed 29 studies of vocabulary learning, and concluded that corpus use benefitted in-depth and long-term vocabulary acquisition; moreover, learners’ corpus consultations promoted the self-driven construction of vocabulary knowledge. However, the use of CT to teach the use of reporting verbs for citations in academic writing has received little attention.

Reporting verbs help to attribute content to another source, and to present, evaluate and comment on scholars’ claims, as well as to indicate whether these claims should be accepted (Hyland 2002). A review of the literature revealed several of the ESL learners’ challenges when reporting citations, including selecting appropriate tenses, learning sentence patterns to integrate citations, and understanding the evaluative nature and discourse functions of individual words (e.g., Huang 2022; Kwon et al. 2018; Lee et al. 2018; Manan and Noorizah 2014). For example, Lee et al. (2018) revealed that L2 learners tend to overuse the pattern according to, instead of the formal and professional reporting structure of X + verb + that clause (e.g., Bigot (1992) states that … ). Huang (2022) found that undergraduate students overuse the verbs to express thinking activities (e.g., think, realise), and underuse the verbs for evidence-based argumentation (e.g., identify), thereby rendering their writing less convincing.

To assist learners to make appropriate lexical, syntactic and rhetorical choices with regard to reporting verbs, Callies (2016) suggested designing classroom activities that allowed learners to explore high-frequency reporting verbs independently by examining expert and learner corpora. In this paper, we discuss the implementation stage by introducing CT, including both expert and learner corpora, to pre-service teachers to teach them reporting verbs.

4.1 Participants and research context

In total, 88 Year 1 and Year 2 undergraduate ESL pre-service teachers of English Language Education (ELE) attending a university in Chinese Hong Kong were recruited for this study. They mainly came from Chinese Hong Kong and the mainland of China, with the age range of 17–20; some may have had one-year English tutoring experience. The majority of the participants were female students since ELE students were primarily female in that university. Their writing skills were equivalent to IELTS 5.5 or above, meeting the admission requirement of the university. The participants were enrolled in a university vocabulary course that focused on learning the meanings and contextual usage of words. Their course coordinator allocated the students to three classes, with 31, 33 and 24 students, respectively. As mentioned by their course coordinator and confirmed during the tutorial classes and interviews, they lacked prior knowledge about the use of reporting verbs for citations, a crucial linguistic feature necessary for completing many of their course assignments.

4.2 Designing features of Parallel EAP Corpora

This study incorporates recommendations from Granger and Tribble (2014) and Gilquin (2022) on integrating learner corpora into CT pedagogy. Their recommendations include representing students from a specific educational background, establishing norms for relevant genres or subject areas, and focusing on specific types of texts. Therefore, this study used the free-access Parallel EAP Corpora (Ma 2014) (interface shown in Figure 2), as it offers three unique functions corresponding to their three recommendations. Firstly, the Corpus function enables users to search for words in academic texts written not only by experienced researchers, but also by L2 university students who share the similar educational background of the participants of this study. Secondly, by searching using the Subject function, participants can choose the genre English Language Teaching (ELT) Research, which is relevant to their academic domain. Thirdly, the Section function assists users to focus on the Literature Review section, which typically contains the highest number and density of citations in a paper. In addition, Parallel EAP Corpora also involve the basic functions of corpus tools, including the generation of concordance lines and collocation lists, thus enabling detailed analyses of the lexico-grammatical features of the target words.

Figure 2:

Interface of Parallel EAP Corpora.

4.3 The instructional design

A two-hour online tutorial via Zoom was designed and conducted by the first author, once for each of the three classes of pre-service teachers. Guided by the proposed theory-supported corpus pedagogy, the instructional design of the learning activities was primarily designed by the authors following the guidance provided in CT research on academic writing and development of corpus literacy (e.g., Ackerley 2017; Callies 2016; Cotos 2014; Gaskell and Cobb 2004; Johansson 2009; Ma 2024; Ma and Mei 2021; Ma et al. 2023). In addition, suggestions of the course coordinator were incorporated to make the learning materials more suitable for the pre-service teacher participants in learning reporting verbs. Figure 3 presents the sequential three-stage training based on socio-cultural and constructivist learning theories, and elaborates the integration of both the expert and the learner corpora in the Parallel EAP Corpora to enhance participants’ noticing of the target items of reporting verbs.

Figure 3:

The three-stage training model.

Throughout the tutorial, participants were asked to directly work with Parallel EAP Corpora to learn the use of reporting verbs and develop all the components of corpus literacy, including 1) understanding of corpora, 2) corpora search skills, 3) analysis of corpus data, 4) advantages of corpora, and 5) limitations of corpora. The three most frequent reporting verbs in ELT research, find, suggest and show (Yan and Ma in progress), were incorporated sequentially into the three stages of teacher-led interventions, peer collaborations and individual investigations, to give participants insights into tense and voice choices, sentence patterns, and the discourse purposes underlying the use of verbs for citations when writing a literature review.

In the first stage, the instructor introduced reporting verbs and Parallel EAP Corpora. Since the participants had little prior knowledge of corpus, the instructor elaborated several key terms of corpus in this stage, including definitions of corpus, concordance lines, keyword in context, word tags, etc., to facilitate participants’ understanding of corpora. This stage also involved teachers’ clear demonstration and step-by-step guidance of searching the word find and analyse its usages based on the concordance lines and collocation lists in expert and learner genres, through which participants can develop the next two components of corpus literacy, i.e., corpus search skills and analysis of corpus data.

The participants then engaged in collaborative activities to reinforce their understanding of reporting verbs and abilities of corpus search and analysis. They worked with their groupmates to search the word suggest, and analyse the concordance lines and collocation list of suggest to notice the usages of reporting verbs and compare the different usages between expert and learner writing to enhance their notice of the target word. This stage enhanced participants’ development of corpus search skills and abilities to analyse corpus data through collaborative activities, and encouraged them to discuss the benefits and limitations of corpus with their peers.

In the third stage, participants were allowed to independently search the word show through Parallel EAP Corpora and analyse its usages in expert writing and learner writing. This stage helped participants to examine whether they had developed appropriate understanding of corpora and sufficient corpus search and analysis skills to implement independent learning activities. Clear guiding questions were also provided in case some participants might encounter any challenges in searching and analysing the target word. Through independent interaction with corpus, they were further encouraged to reflect on the benefits and limitations of corpus.

Furthermore, after the tutorial, supporting materials were also distributed to participants, which consists of instructions about the use of Parallel EAP Corpora to learn other reporting verbs, such as investigate, note, and think, and the use of using other popular corpora and corpus tools, such as COCA and AntConc.

4.4 Data collection and analysis

Five types of data (Table 1) were collected in the sequence of

1. classroom observations,

2. Likert-scale questionnaires,

3. open-ended question,

4. interview transcripts, and

5. tests concerning the reporting of citations.

A trained research assistant conducted the classroom observations during the online tutorials to document how the learning activities facilitated the pre-service teachers’ learning progress. Using the similar template in Ma and Rui et al. (2024), the research assistant independently observed all the three tutorials, each of which lasts for 2 h.

5.1 Pre-service teachers’ learning outcomes

Two major learning outcomes were summarised: (1) using reporting verbs for citations, and (2) acquiring necessary corpus literacy.

5.2 Pre-service teachers’ evaluations of the theory-supported corpus pedagogy

Thematic analysis of the qualitative data, including classroom observations, open-ended data and interview transcripts, revealed that the pre-service teachers enjoyed these learning activities under the theory-supported corpus pedagogy and assessed them positively. The analysis identified 4 main themes: (1) teacher scaffolding, (2) peer collaboration, (3) independent corpus search, and (4) the usefulness of Parallel EAP Corpora.

5.2.1 Teacher scaffolding

Of the 35 participants who responded to the open-ended question about their favourite tutorial activities, 16 mentioned teacher scaffolding. They appreciated the “demonstration of using corpora”, “clear instructions and exercises to work on” and the “steps in using Parallel EAP Corpora shown in the tutorial”.

The participants may have enjoyed teacher scaffolding because it addressed their difficulties in using Parallel EAP Corpora. The research assistant observed that “some students found it difficult to use Parallel EAP Corpora and encountered many challenges” at the beginning of the tutorial. The instructor adopted the strategy of “demonstrating the search processes of the word find and providing clear search steps to the students” in response, as shown in Figure 4.

Figure 4:

A screenshot that the research assistant captured during the third tutorial.

Similarly, the interview data showed the participants’ concerns that using Parallel EAP Corpora might be challenging. Of note, S6 expressed that “the practical step-by-step guidance and the very specific practice offered in the tutorial helped us to understand how to use the corpus, and I would like to apply these strategies in my future teaching” (S6).

5.3 Pre-service teachers’ intentions regarding the use of CT

The results for Items 7–9 (Table 7) showcase pre-service teachers’ three general intentions regarding the use of corpora, which were for future learning (M = 3.12, SD = 0.59), research (M = 3.10, SD = 0.54) and teaching (M = 3.16, SD = 0.55) purposes. These results highlight the tremendous value achieved through a single tutorial, which increased the pre-service teachers’ interest and demonstrated that corpora can be utilised not only for learning purposes, but also for conducting language analyses and facilitating language teaching.

The thematic analysis of the interview transcripts further revealed three specific intentions regarding the use of CT:

1. improving vocabulary learning and academic writing,

2. conducting research to complete course assignments, and

3. facilitating future teaching.

The participants also mentioned their concerns about the use of CT in learning and future teaching, and expressed a strong desire for more CT tutorials.

6.1 Effectiveness for learning the use of reporting verbs and developing corpus literacy

The tutorial addressed pre-service ESL teachers’ learning needs regarding the use of reporting verbs for citations, following the long-standing recommendation that source work should be the focus of writing instruction for early year undergraduates (Kwon et al. 2018). This study demonstrated the effectiveness of employing both expert/native and learner corpora to learn reporting verbs, and reinforced what established language learning theories can offer to CT-based pedagogical practices.

Furthermore, the participants developed their corpus literacy through participating in the CT training supported by the theory-supported corpus pedagogy. They not only learned several basic concepts of corpora and how to conduct search functions of Parallel EAP Corpora, but also were equipped with the skills of analysing corpus data and developed a deeper understanding of the advantages and limitations of CT. As shown in Ma et al. (2023), all of these are crucial aspects of developing corpus literacy, which further influences pre-service teachers’ future intention to adopt CT in classroom teaching.

6.2 Significance of the theory-supported corpus pedagogy

Corpus linguists propose that constructivist learning, socio-cultural theory and the noticing hypothesis can be linked to CT to make students’ learning more effective (Crosthwaite and Boulton in press; Flowerdew 2015; O’Keeffe 2021). In this study, we successfully integrated these language learning theories into corpus-based learning designs, marking a significant advancement over many prior studies in corpus linguistics that often lacked robust, theory-supported pedagogical frameworks (Boulton and Vyatkina 2021; O’Keeffe 2021). This integration highlights our contribution to enhancing the theoretical grounding and practical application of corpus-based language education in several ways. First, our study shows that the social constructivist learning theory can help participants learn corpus use and target items in a sequential and well-organised process through teacher-led intervention, peer collaboration, and independent learning. Our research demonstrates how active learning combines the advantages of constructivist learning – which focuses on students’ active engagement in inferencing and hypothesizing, as outlined by Bada and Olusegun (2015) and Brown and Palincsar (1989) – with the principles of socio-cultural theory as well as noticing hypothesis. The sociocultural theory emphasizes the critical role of teachers and peers in mediating and facilitating the construction of knowledge, as described by Jaramillo (1996). Together, these theories underpin our approach to integrating interactive, collaborative elements into the learning process, enhancing student learning outcomes. Second, through combining expert and learner corpus data into the activities under social cultural and constructivist learning theories, it provides evidence of how learner corpus data can be integrated for noticing enhancement, as claimed by Schmidt (2010) that providing data generated by learners themselves can augment them to notice the gaps of their own language use. Our results further revealed that the learners evaluated the experience positively and engaged actively in the learning activities, thus highlighting the significance of this theory-supported corpus pedagogy.

The successful implementation of the theory-supported corpus pedagogy is firmly grounded in the data and functionalities provided by the Parallel EAP Corpora. This resource offers abundant writing samples of literature reviews, enabling participants to explore the use of reporting verbs for citations in the Literature Review section. Its user-friendly interface minimizes the technical demands on teachers when preparing clear scaffolding materials and facilitates students’ ability to conduct independent corpus searches and analyses efficiently, thereby helping them to construct the language use patterns. Additionally, Parallel EAP Corpora allows users to compare both expert and learner academic writing samples, which fosters their noticing of the target items. Given the positive feedback from the participants, it is recommended that future EAP studies and practices incorporate Parallel EAP Corpora or design similar corpora to enhance students’ advanced academic writing skills.

6.3 The benefits of introducing CT to pre-service teachers

After recognizing the benefits of Corpus Technology (CT) for their own learning, acquiring the necessary corpus literacy, and observing the importance of theory-supported corpus pedagogy, the pre-service teachers began to critically reflect on how and whether CT could be implemented in classroom teaching to enhance their future students’ language learning. Römer (2006) suggests that incorporating CT into more educational contexts should focus on teacher education, and should target initial teacher training at universities, practicing teachers and advanced learners. However, despite this on-going call, CT has had limited influence on language teachers’ classroom practices (Boulton 2017; Chambers 2019; Ma et al. 2022). The findings of this study provide new insights to address this issue by demonstrating that integrating CT into pre-service teachers’ language learning can also serve the purposes of pre-service teacher training.

Furthermore, this study revealed that CT could facilitate pre-service teachers’ academic development as novice researchers by expanding their methodological toolbox for language research (Paquot and Gries 2021), thus enabling them to contribute to the fields of applied linguistics and language education (Cheng et al. 2003; Mussetta and Vartalatis 2018).

In summary, many pre-service ESL teachers have multifaceted identities as language learners, future teachers and novice researchers, shaping their experiences and interactions with CT. They are also critical thinkers who reflect on the extent to which CT can support their own language learning and their future teaching in different contexts. Focusing on these pre-service teachers facilitates the future development of CT in the domains of language research and education.