Introduction

Corpus might have a very general definition as the one given by The Experts Advisory Group on Language Engineering Standards (EAGLES) mentioned by Meyer (2004), which states that corpus might be any text type collection as newspaper, poetry, drama, dictionaries, etc. Perhaps, linguists analysing corpus prefer a more restricted definition, referring to a collection of text used for linguistic analysis. Collins COBUILD English Language Dictionary, published in 1987, was the first corpus-based dictionary and the one that brought corpora alive for English teachers (Grabrielatos, 2005). Since then, corpus-based products have grown exponentially, not only teaching activities, products related to this area as dictionaries or frequent word lists, but also linguistic products as well.

Besides, some different corpora definitions there exit different corpora. They differ in shapes and size because they are built for different purposes. “There are two philosophies behind their design, leading to the distinction between reference and monitor corpora…Another design-related distinction is whether a corpus contain whole texts, or merely samples of a specified length” (Gabrielatos, 2005, p. 3).

• Reference corpora: It has a fixed size (e.g., The British National Corpus).
• Monitor corpora: It is expandable (e.g., The Bank of English)

In terms of content, corpora can be general or specialized:

• General corpora: Reflect a specific language or variety in all its contexts. (e.g., the American National Corpus)
• Specialized corpora: Focus on specific contexts and users (e.g., Michigan Corpus of Academic Spoken English)

Corpora is relevant to language teaching in many ways since concordance offers realism and relevance. It is a common practice for language teachers to create inauthentic activities which are correct in the target language, but might not exist in a real context. Concordance-based materials can offer real and relevant texts (Ma, 1993). In addition, “corpora of language teaching coursebooks enable the examination of the language to which learners are exposed” (Gabrielatos, 2005, p. 4).

The importance of a language corpus is that it contains examples of realm up-to-date language use. In the past, researchers used the language they heard or read from people in their circle of friends as examples. These examples were intuitive and they were often correct or nearly correct, as stated by Fox (1998), but they were not always reliable or related to what people, in general, were really saying. Currently, teachers still use introspection to create their activities and while many grammar structures are simplified and some exceptions are presented, many other uses of the language are ignored.

Corpus-based research has also revealed that many common language patterns are missing in ELT materials. Corpora help to analyse the language forms that are used by native speakers in which context and frequency give language learners a more reliable model to follow. For example,

The transitive use of ‘give’ in clauses such as ‘she gave him a really lovely smile’, and ‘he gave an extremely boring talk’ is so common that many native speakers hardly even notice it. The meaning of ‘give’ here is easily intelligible to learners, but what they might not realize is how important this particular structure pattern is…it allows us to focus on the event rather on the action (Fox, 1998, p. 27).

Concordance Method for Corpora Searching

According to Hunston (2002), “Producing concordance lines is the most basic way of processing corpus information” (p. 38). Concordance lines are “all the instances of a word or phrase in the corpus.” (Bennet, 2010, p. 17) To obtain concordance lines from a corpus, a concordancer is needed. This is a program that searches a word or phrase in a corpus and then presents it on the computer screen with the words that come before and after it. The word that is searched is called node word.

The concordance lines produced in the search can be sorted in alphabetical order using the word to the left or the right side of the node word for the sorting. This sorting makes it easy to analyse the results and find patterns on how that node word is used (See Appendix 1). Understanding concordance lines needs some practice and as mentioned by Breyer (2009), it is important that teachers receive corpora and concordance teaching during their initial training programs. She expressed that since corpora is still not included in the curricula, it is the teachers’ decision to incorporate corpora-based activities in their teaching at present time.

Breyer (2009) pointed out some key factors that influence teachers’ decision on implementing corpora into their teaching:

• Motivation.
• Availability of materials.
• Possession of adequate skills to teach with corpora.

Throughout her investigation, Breyer (2009) concluded that the best moment to include corpus training was during initial teacher education; “in this context, the potential exists to integrate it meaningfully into their training and to devote sufficient time to the subject to allow for in-depth exploration” (p. 4). It is important though, to teach the teacher learners their role as teachers when applying corpora in their classes, this is not the typical role a teacher has. When using corpora, the teacher “has to learn to become a director and coordinator of student-initiated research” (Johns, as cited by Breyer, 2009. p. 4).

The results presented in Breyer’s (2009) case study in which student teachers initiated their corpora training, it can be observed that the student teachers were concerned at the beginning about how language teaching had become so simplistic while trying to make language learning easier for learners. They felt that process was not giving learners the correct picture, as demonstrated in the following comments:

I suppose it could be a problem if they just learn the general rules and then, eventually, are confronted with sentences which do not fit into the system they were taught. I assume this is very confusing and it is hard to look beyond the rules, which one has once learned. (Breyer, 2009, p. 10).

The point is, however, that over-simplification leads to incorrect portraying of authentic language use…Pupils are taught only a part of the rule in a bid to keep it straightforward and simple (Breyer, 2009. p. 10).

Later, while student teachers were advancing in their corpora training, they came up with some ideas on how to incorporate the use of corpora. As stated by Breyer (2009), student teachers suggested the following procedure:

• Teach the basic language rules for beginners.
• Once language learners had learned basic language rules, they can discover extended rules on their own using a concordancer. Data-driven learning is a good option.

This study gave a clear view about how student teachers became aware of the importance and usefulness of corpora in their teaching, and demonstrated it could be useful to introduce corpora training into teaching education.

What to observe in concordance results

Once teachers have been trained on how to use corpora, they will be able to make the correct observations; there are different types of observations that help to obtain useful information from concordance:

1. Observing the central and typical.
2. Observing meaning distinction.
3. Observing meaning and pattern.
4. Observing detail.

Typical: “Might be used to describe the most frequent meaning or collocates or phraseology of an individual word or phrase” (Hunston, 2002, p. 42). For example, if you search for the concordance lines for the word “recipe” you would be able to observe that the typical meaning for this word is metaphoric, not a literal recipe, as in the following concordance line:

“to god material, and you have a recipe for serious success. But is not a” (2002, p. 43).

Centrality: “Can be applied to categories of things rather than individual words” (Hunston, 2002, p. 43). For example, talking about the uses given to the present progressive; it is known that it is used to talk about the present, the future, or not specific time; however, it is more frequently used to talk about the present; therefore, she stated that the central use for present progressive is to talk about the present time.

Among these two concepts, surges a third one: prototypical; it is used to indicate a common usage that can be perceived as typical, but is not necessarily the most frequent. It has been demonstrated that prototypical examples are frequently used in language coursebooks (Hunston, 2002). An explicit example is the way that comparatives are presented, where the grammar form indicates that the word than is part of a comparative sentence, but concordance has demonstrated that typically it is not used as in the following sentences: “a much larger plan, or their larger but poorer northern neighbours” (p. 44). In these previous sentences, the comparison is implicit.

Meaning distinction: There are words with similar meaning, but they cannot replace one another; they are called near-synonyms. Concordance lines can help to clarify their different meanings.

Meaning and pattern: “Meaning of a word is closely related with its co-text” (Hunston, 2002, p. 46). Its meaning depends on the patterns or phraseologies where they occur. If the pattern is distinguished, then the meanings are as well. This applies when searching for individual words. There is also the option to obtain meaning for words with similar meaning that share patterns. This last option can be done by “searching for phrases and noting the words that frequently occur within it” (p. 48).

A variation of searching phrases to obtain meaning is the concept of frames; It is a sequence of, usually three words in which the first and last are fixed, but the middle word is not. For example:

too + adjective+ to --> too young to, too easy to, etc.

Detail: Observing details helps to lead to more specific observations about the specific behaviour of individual words.

Concordance in the Language Classroom

As mentioned before, concordance lines can be exploited in the language classroom to help make language learners aware of how language is used. This has been demonstrated by previous studies which show how the concordance approach to teaching vocabulary is more effective than traditional methods, as demonstrated by Gan et al. (1996).

In their study, Kaur and Hegelheimer (2005) analysed the effectiveness of using an online concordance and dictionary in acquiring new vocabulary, and comparing this procedure with using only an online dictionary. The study focused on helping learners to improve their academic writing skill by increasing their vocabulary. This study highlighted the importance of having extensive vocabulary knowledge for ESL learners in order to produce the types of writing expected from them in an academic setting, and re-marked how difficult it is for ESL learners to acquire vocabulary due to the limited direct contact they have with the target language. Therefore, ESL learners are limited to produce language due to their poor lexical capability. In this study, Kaur and Hegelheimer, worked with a group of learners who were enrolled in a writing course for ESL undergraduates after taking the English Placement Test (EPT). Their writing evaluation showed deficiency in their composition writing. A group of language professors selected thirty words from the Academic Word List (Coxhead, 2000) for the learners to use to develop a specific writing task. The concordancer selected was Tom Cobbs’s Compleat Lexical Tutor and the Dictionary.com was the online dictionary used. Learners were indistinctly separated into two groups; one group could use the concordancer and online dictionary and the other one could not.

The learners went through the following procedure:

1. Answering a questionnaire to identify their language.
2. Exposure to the use of a concordance program.
3. Answering a pre-test to evaluate their receptive academic vocabulary knowledge.
4. Solving a cloze activity.
5. Solving a sentence-building task for learners to demonstrate their understanding of the meaning.
6. Trying to collocate the target vocabulary to observe if they could use it productively.
7. Solving a writing task where learners could demonstrate their productive knowledge.
8. Answering a post-questionnaire to obtain opinions from learners about how positive the experience with concordance was.

After this procedure, it was demonstrated that those learners who had access to the concordancer and online dictionary had improved their performance on using the target words correctly than those who did not have access to these tools. In addition, it was also showed that both groups of learners used the target vocabulary in their writing task. Nevertheless, the ones working with the concordancer and online dictionary made more attempts and had more correct words compared to the learners who did not use these tools.

Therefore, it is useful to encourage language teachers to develop corpus-designed activities, to help learners interact with native-like language use and learn from this interaction.

A framework for creating corpus-designed activities

Direct use of corpora for teaching includes three different focuses according to Leech (1997):

• Teaching about: Teaching corpus linguistic as an academic subject.
• Teaching to exploit: Teaching learners how to use corpora by themselves (hands-on). The learning activity may become learner-centred.
• Exploiting to teach: Using a corpus-based approach to teaching language and linguistics courses.

As McEnery and Xiao (2011) expressed, teaching about and exploiting to teach are associated with students of linguistics and language programs, while teaching to exploit is related to students of all subjects including language study and learning who are expected to derive benefit from Data-driven learning (DDL) and Discovery Learning.

There are different approaches to creating corpus-designed activities like the three Is approach (Illustration, Interaction, and Induction) mentioned by McEnery and Xiao (2011), where illustration refers to looking at real data, interaction is discussing and sharing opinions and observations, finally, induction refers to making one’s own rule for a particular feature of the language. This approach has the drawback of provoking partial and incomplete generalizations from limited data “as a stage on the way towards a satisfactory rule” (p. 370).

Another approach is the Data-driven Learning (DDL) suggested by Johns (1991). This approach is learner-centered, and guides the learners in an autonomous learning process where the learner becomes the researcher. Since making a correct interpretation of corpora results is not an easy task, McEnery and Xiao (2011) mentioned that teachers' guidance and pedagogical mediation is needed. Therefore, as mentioned before, corpus training for language teachers is necessary.

The DDL (Data Driven Learning) approach includes three stages of inductive reasoning: observation of concordance evidence, classification of salient features, and generalizations of rules, similar to the ‘Three Is’ approach (Johns, 1991).

Bennet (2010) suggested the following framework which offers seven clear and easy-to-follow steps. The first five steps are for the teacher to identify the target language to work with, and the type of corpus to be used. The last two steps are related to the activity creation where the ‘three Is’ or ‘DDL’ approach might be used.

Figure 1. A framework for developing corpus-designed activities (Adapted from Bennet, 2010, p. 19)
A similar framework, developed in six steps, is mentioned by Bardovi-Harlig et al. (2015):

1. Selecting the corpus
2. Identifying expressions
3. Extracting examples
4. Preparing corpus excerpts for teaching
5. Developing noticing activities
6. Developing production activities

This framework was used to analyse expressions since the focus of the article was on pragmatic routines (expressions), such as Thank you so much and That’d be great! for thanking and Nice to meet you for introductions. It can be observed, in this second framework that the corpus is selected before identifying the expression (research question step in Bennet’s (2010) framework). Once these two steps covered the following steps, even though expressed differently, are similar in function: They use the selected corpora, analyse the results, adapt the aspects to facilitate the target language (TL) coverage, create the corpus-based activities for learners to notice the TL and then provide learners with the opportunity to experience/produce it.

Application in a specific context

As an example, this paper presents an activity created for intermediate level students in a language school, which belongs to a public university in Mexico. These students are in their final semester of a three-year-long English program.

This activity was designed to clarify the use of the word miss, since the students seemed to be confused with its use and meanings. The activity development follows Bennet’s framework. It was selected since it guides the researcher in a detailed form through the entire process; the steps to follow present a logical sequence.

Figure 2. Process to create a corpus-designed activity

Conclusion

Corpus-based activities give students contact with real use of language, which prepares them for real-world interaction. These activities could be as varied as the teacher’s imagination is. Students learn to observe patterns, and it develops their language awareness helping them to be independent learners which increases the probabilities of their success in their language use. There is one important aspect to consider; teachers need to be trained in corpus use, searching techniques and corpus-based activities creation since their first encounter with corpora might be confusing.

References

Bardovi-Harlig, K., Mossman, S., & Vellenga, H. E. (2015). Developing corpus-based material to teach pragmatic routines. TESOL Journal, 6(3), 499-526. https://doi.org/10.1002/tesj.177

Bennet, G. R. (2010). Using corpora in the language learning classroom: Corpus linguistics for teachers. University of Michigan Press.

Breyer, Y. A. (2009). Learning and teaching with corpora: Reflection by student teachers. Computer Assisted Language Learning, 22(2), 153-172. https://doi.org/10.1080/09588220902778328

Coxhead, A. (2000). A new academic word list. TESOL Quarterly, 34(2), 213-238. https://doi.org/10.2307/3587951

Fox, G. (1998). Using corpus data in the classroom. Cambridge University Press.

Gabrielatos, C. (2005). Corpora and language teaching: just a fling or wedding bells? TESL EJ 8(4), 1-35. https://www.tesl-ej.org/wordpress/issues/volume8/ej32/ej32a1

Gan, S.-L., Low, F., & Yaakub, N. F. bte (1996). Modeling teaching with a computer-based concordancer in a TESL preservice teacher education program. Journal of Computing in Teacher Education, 12(4), 28-32. https://doi.org/10.1080/10402454.1996.10784301

Hunston. S. (2002). Corpora in applied linguistics. Cambridge University Press.

Johns, T. (1991) Should you be persuaded: Two samples of data-driven learning materials. English Language Research, 1-16.

Kaur, J., & Hegelheimer, V. (2005). ESL students' use of concordance in the transfer of academic word knowledge: An exploratory study. Computer Assisted Language Learning, 18(4), 287-310. https:/doi.org/10.1080/09588220500280412

Leech, G. (1997). Teaching and language corpora: A convergence. Routledge.

Ma, B. K. C. (1993). Small-corpora concordancing in ESL teaching and learning. Hong Kong Papers in Linguistics and Language Teaching,16, 11-30. https://files.eric.ed.gov/fulltext/ED365119.pdf

McEnery, T., & Xiao, R. (2011). What corpora can offer in language teaching and learning. In E. Hinkel (Ed.), Handbook of research in second language teaching and learning (pp. 364-380). Routledge.