What we dare hope for

Talking to colleagues in the field of linguistics and language education and reading the literature, there is clearly a groundswell of enthusiasm for a more widespread use of corpora by language teachers - the term Data Driven Learning (DDL) has long been in vogue to describe the approach. Paradoxically, many of these professionals do not use them "at the chalkface" themselves for a variety of reasons.

The problems with implementing DDL in the classroom are well-documented and well-understood. The list includes:

• • lack of equipment

  • lack of training

  • lack of published resources

  • access to corpora and concordances

  • unfriendly nature of concordancing software

  • radical departure from normal teaching styles

  • confidence in interpreting data, esp on the fly in front of students

  • corpora contain too much attested language (an ironic comment to be sure)

  • teaching styles and learning styles

Knowing this doesn't stop any of us wishing to see corpora being consulted in:

• • course design

  • lesson and resource preparation

  • during a lesson

At this stage it is difficult to impose corpus use on teachers as there is not enough research demonstrating its effectiveness. See Alex Boulton's article, But where's the proof? This begs the question how linguists and language education professionals, of which I am one, have the gall to impose this on others. But let's not go there for the minute.

I have some non-imperical evidence from the Czech students I trained to consult corpora in the process of teaching them academic English - some unedited extracts of their testaments can be found here from which several things emerge:

• • despite their praise, there are still plenty of errors in their English

  • they use advanced terminology for describing language per se, not all of which they learnt in my classes

  • language is not their field - they are doctoral students of computer science. This may predispose them to using computers and to analysing data

  • they are acutely aware of the problems they have when writing texts in English

  • they take whatever good they can get from their corpus analysis but consult other sources as well.

Modern language pedagogy places great emphasis on vocabulary, collocates and multi-word units. The pioneering pedagogical work of the Cobuild Dictionary and Michael Lewis' Lexical Approach in the early 1990s set this in motion and it remains the core of much ELT pedagogy today.

So, let's consider a plan of attack that could see the widespread use of corpora by language teachers. The professionals mentioned at the top of this entry need to be conscripted, which given their motivation, would not be too difficult. Essentially, we have to train the next generation of teachers which means that that the process must begin in teacher preparation institutions.

I would like to see a compulsory corpus course in the first year of language degree programs so that throughout all their language studies, which includes language, literature and cultural studies, the students would be equipped to research any language point being studied. Later, those who choose to specialise in teaching would take such language research for granted and would carry it over into their teaching.

Sounds simple, doesn't it? There are plenty of obstacles, of course. For example, not all language departments have staff who could do such training or have access to software adequate to the task.

I didn't start this posting as an advertisement for my own e-learning course, rather it was motivated by the dissertation outline of one of my masters students. However, I am writing an e-learning course for language teacher trainees: Moodle is the e-learning software and the corpus software the course teaches is the Sketch Engine, since this is created at Masaryk University where I work. When it is fully written and pilotted, which should be by the end of January 2009, it will be available to all (1).

Other obstacles include those alluded to above such as the confidence non-native speakers have in drawing reliable conclusions from data, faith in the probablistic nature of language - a leap of faith for many, "time well-spent" issues, too much unfamiliar language in corpora. In my experience, these issues tend to iron themselves out with frequent use and encouragement. An understanding and application of David Nunan's maxim adapt the task, not the language is always helpful.

The oft-cited obstacle, access to corpora and concordances, is simply not true. On my web-based corpora page, there are enough resources to keep a language student and teacher busy for an age. Most of them are free, but those which are not are worth paying for in my opinion - we do pay for books and DVDs and journal subscriptions, etc.

So let us hope that teacher training institutions will start preparing their trainees to be confident, independent language learners and in the future, teachers who share the new understanding of language and language learning that emerges from a data-driven approach.

I originally posted this article on my blog.