Getting students to use corpora themselves

"'Tis all in pieces, all coherence gone"
(John Donne, The Anatomy of the World 1621)
Nick Pringle and Kris van de Poel, University of Antwerp, Belgium

Background

In "traditional" foreign/second language teaching it is accepted that practice in production is an important part of target language (TL) acquisition, and that it is necessary for the gradual mastery of explicitly learned TL knowledge. Thus, language production enhances the automisation process of language learning. But how can this production be made more accurate?

A 'problem' area for intermediate/advanced learners: chunks

One of the things that I am always struck by when teaching English to both intermediate and advanced level students is that the majority of their mistakes/errors are lexical rather than grammatical. Moreover, they seem to find it difficult to raise their level of proficiency. As a result, I often try to impress upon learners the importance of being able to break language down into different 'lexical items', particularly multi-word items (or chunks), in order to help them build up their mental lexicons.

However, I often encounter two main obstacles to this process of raising lexical awareness. First of all, some students are extremely concerned about avoiding grammatical errors, even if these do not cause problems of meaning and could be avoided if they knew more lexical items. The second problem is probably best summed up by John Donne in his poem The Anatomy of the World, when in 1621 he lamented that "'Tis all in pieces, all coherence gone" as scientific progress and social change began to sweep the old certainties away. The same is true, or seems to be true, with any shift away from time spent explaining structure and rules, and the comforts that learning them may seem to offer, to a more memory-based and never-ending system based on noticing and learning large quantities of chunks - all coherence indeed seems to be gone. I should point out that in most cases the students I have taught have a very strong grasp of linguistic rules and structure and seem to reinforce thei r liking by focusing even more on them.

Approaches to learning lexical chunks

To illustrate the importance of noticing lexical chunks, or of what Michael Lewis has called 'the linguistic wrapping in which the message is delivered' (Lewis 2000) I tend to focus on three key areas, all of which involve the use of corpora: 1) presenting learners with negative evidence (which is often more effective than giving five good examples), 2) letting learners know that computer concordance programs exist by giving them some useful websites (the medium often draws them towards the topic), and 3) encouraging learners to compile an electronic/paper corpus of their own (a personalisation of language learning).

1) A natural response is for students to translate 'chunk for chunk', often creating false friends, if they do not know a particular lexical phrase in English. In other words, they retrieve the appropriate L1 lexical unit that they naturally associate with a particular situation and then translate it word-for-word into English - a natural strategy for anyone who has never encountered the appropriate collocation in English, or who has failed to notice and remember it. Thus, native speakers of Dutch often translate the phrase 'volgens mij' into 'according to me' to inform the listener that they want to express a personal opinion. When faced with a phrase like this - that sounds possible but not probable - I often like to consult a corpus rather than simply relying on my own database of prefabricated phrases (i.e. my memory). In this particular case a simple search of the British National Corpus (BNC) revealed only 6 entries of 'according to me'. Students, even adult learners an d lecturers in language training programmes, appreciate being presented with negative (contrastive) evidence of collocations that they use but which do not exist (so-called 'blocked collocations'), or collocations which are highly improbable.

2) As mentioned above, I like to make students aware of the existence of computer concordance programs and of how they can be used to retrieve lots of real language data. In my experience, students tend to warm to the idea of using corpora for exploring the limits of grammatical collocations, such as to check which preposition collocates with a particular lexical item (difficult to look up in a dictionary). Thus, corpus analysis may reveal to students the importance of collocation and colligation. On the other hand, a corpus can be a rather blunt tool to put into the hands of a student who may not have the time or inclination to use one - at least until more student-friendly versions become available online, which categorise the lexicon in more easily digestible ways for students.

3) Students seem best able to see different lexical items in a text by getting them to distinguish between pragmatic (and often modal) expressions and high-content collocations, (particularly high-frequency verb + noun collocations and noun phrases). Breaking down a simple sentence such as 'Would you like a cup of tea?' into two parts - would you like and a cup of tea - clearly illustrates how the language can be chopped up and rearranged. For example, the usefulness of the combination would you like when students want to offer something becomes clear. To build on this I like to give students a few short texts to ask them where they think the articles came from and whether or not they are useful to anyone learning English. This is a particularly good exercise to get students to use or reuse phrases for making guesses and stating opinions. In addition, it also gets students to think on different cognitive levels because they obviously have to read for meaning as well as form and poss ible origin. A question which I always ask is, if they can find any word combinations that they recognise or that frequently occur throughout a text. If they are still unsure I mention a verb + noun combination that occurs frequently throughout the text (e.g. watching + television). This can be particularly striking when the article is short. In addition, I sometimes point out a few common collocations that they are already familiar with such as 'junk food'. To develop lexical awareness still further I encourage students to read through articles related to their field of interest for form as well as meaning. The purpose of reading for form, I explain, is so that they notice lexical items that can easily be missed, such as the more pragmatic language of sentence heads and sentence adverbs that often seem to receive less attention from the reader than high-content collocations. If learners see the value of doing this and are encouraged to do it on a regular basis, it could help them to build up their own pers onal 'paper' corpus.
I went through several articles myself to illustrate to my students just how easy it is to pick out lexical items that are often used in academic articles in English. Here are just a few which I used to write this paper, and of course, there is absolutely no reason why students cannot do the same:

The value of such a list to even advanced learners is that they would have more thinking time to concentrate on more subject-specific terminology, argumentation, etc. And though more focus on corpus data and lexis could possibly lead to some student frustration at the apparent lack of rules and the seeming unlimited possibilities, personal corpora for specific purposes could help restore coherence in the minds of learners, even if the language is 'all in pieces' - chunks.