Jamie Keddie examines conventional corpora – the databases of language that have been compiled specifically for the purpose of linguistic investigation – to see how they can be used to strengthen students’ understanding of the importance of collocation when learning a new language.

Photo of someone looking something up on the internet.

Source: PeopleImages, Getty Images

Introduction: Three language learning / teaching situations

Situation 1: A Spanish student is telling the rest of the class about the time he 'went on a travel' to Thailand and Cambodia. Following his anecdote, I write his mistake up on the board and ask the rest of the class if they can correct it. They correctly identify the error and change it to 'went on a trip'. As a closing afterthought, I tell everyone that the word 'travel' is a verb and not a noun.

Situation 2: I have a German student called Hans-Peter (name changed to protect identity). Without a doubt, Hans-Peter’s favourite English word is the adjective 'nice' and he likes to use it as much as possible when conversing. As a result, his English can sound a bit odd at times. Take, for example, the day he described an investigative technique that he was working on as 'very nice', rather than 'very effective' (he is a molecular biologist). My early efforts to correct this particular problem, which focused on the inherent meaning of the adjective in question, were largely unsuccessful.

Situation 3: A class of intermediate learners come across the phrasal verb 'to carry out'. They ask me what it means. After a bit of deep thought, the only thing I can say is that it is similar in meaning to the verb 'do', although a bit more formal.

Human dictionaries

As language teachers, we are all expected to be able to offer our learners concise and accurate word definitions upon demand. Turning to a teacher for lexical advice is, of course, a lot more convenient than turning to a dictionary. However, this approach is not without its drawbacks and pitfalls. For example, in one of the situations discussed above, I gave my learners information that was quite simply incorrect; the word travel can certainly be a noun. In the other two situations, my attempts to clarify the items in question were largely unhelpful since I was denying my learners the bigger picture.

Can’t see the wood for the trees

If you can’t see the wood for the trees, you are unable to understand what is important in a situation because you are focusing in on unnecessary details. Consider the following three items:

  1. Travel (noun)
  2. Nice (adjective)
  3. To carry out (verb)

Despite the fact that any competent speaker of English could spend a good amount of time discussing the inherent meanings of such items out of context, we really have to ask ourselves what use this would be in the language classroom. Rather than supplying our students with such improvised explanations, it can be a lot more beneficial to supply them with a basic definition (from a good learner dictionary, for example) and then subject them to multiple examples of the language in use.

The British National Corpus

The British National Corpus (BNC hereafter) can be searched via the English Corpora site:

1) Go to www.english-corpora.org/bnc/.

2) Enter a word or item of your choice (for example: travel, nice, or carry out) in the search box. The word will appear listed with its frequency in the corpus.

3) Click on the word or phrase itself to see a list of random contextualised examples. In most cases, the examples given are extracts of text, usually whole sentences.

What follows is three activities that could be created quickly in response to situations such as those mentioned at the start of this article. Each article exploits the real 'specimens' of language that a corpus search would yield.

Activity 1: Memory game

Situation: Students know that the word 'travel' can be a noun as well as a verb, but don’t understand why it is incorrect to say 'I went on a travel to Thailand.'

1) Type 'travel' into the BNC corpus, print out a selection of the results and hand out copies to students.

2) In pairs or small groups, ask your students to identify ten collocations involving the word 'travel' as a noun. Once they have done this, allow them to share and compare their results.

4) Elicit all results and write them on the board. You might be offered any of the following:

  • A travel agent
  • World travel
  • Travel insurance
  • Air travel
  • A travel card (for London transport)
  • A travel guide
  • Travel expenses
  • Travel arrangements
  • A travel operator
  • Space travel

5) Wipe the board clean, take back all the copies of the corpus search and get your students to write down, from memory, as many collocations that were written on the board as possible.

6) Ask your students if they can work out why it is incorrect to say 'I went on a travel to Thailand' by looking at the corpus search results. They may be able to tell you that 'travel' is an uncountable noun.

Activity 2: Collocation poems

Situation: Students want to know exactly what the adjective 'nice' means. You read them a dictionary definition of the word and point out that the most important factor is to consider the things in life that can be described as 'nice'.

1) Type nice' into the corpus and print out a selection of the search results.

2) Make a few photocopies of the search results for your students and have them identify collocations as in the previous activity. Again, write these on the board. You might come across any of the following:

  • Nice enough
  • Nice to know
  • Quite nice
  • Sounds nice
  • Etc

3) Ask each student to select his or her eight favourite 'nice' collocations. Students should write each collocation on a separate little piece of paper. Have them shuffle these around until a poem is born. Here is one that was prepared earlier:

Nice weather
Nice shot
Nice guy
Be nice

Nice and nasty
Nice to see you
Nice people

Nice and easy

4) Get students to read out their poems to the rest of the class. Make sure they pay attention to rhythm as they do so.

Activity 3: Things to carry out

Situation: Students want to know exactly what the phrasal verb 'carry out' means. Tell them that the important thing to bear in mind is not so much the meaning of the verb, but the things that can be 'carried out'.

1) Type 'carried out' into the corpus and print out a selection of the search results.

2) Once again, make a few photocopies of the search results for your students. Have them identify things which can be carried out and write these on the board. For example:

  • A massacre
  • A procedure
  • A repair
  • A study
  • A survey
  • An analysis
  • An inspection
  • An interview
  • An investigation
  • Building work
  • Medical tests
  • Orders (military)
  • Research
  • Restoration
  • Work

3) If you have computer access in the classroom, direct your students to the following site: http://puzzlemaker.school.discovery.com/

They will be able to create and print off word searches for each other from this site. Get them to choose ten things that can be carried out and incorporate these into their puzzles. If you don’t have internet access in class, you can always prepare word searches for your students before the next lesson.

Anticipated problems

I cannot pretend that corpus searches are to every learner’s taste. In fact, they may be compared to marmite, because students will either love them or hate them. A sea of text with no pictorial elements can be enough to scare some students off, so it may be unfair to persevere with activities such as the ones suggested in this article if they are continually unwelcome. On the other hand, if a learner recognises the value of a corpus search, he or she will be equipped with an invaluable tool for self-study.

Here are some suggestions for happier corpus activities:

1) It cannot be assumed that the meaning of a word or item can be appreciated simply because it is partly contextualised. Many of the examples that will arise from a BNC search will be nothing short of useless. Rather than printing off the results of an entire corpus search, copy and paste the contextualised examples onto a word document. This will allow you to delete the unhelpful ones.

2) By cutting and pasting the corpus examples into a word document, you will also be able to make gentle alterations to any potentially obscure or problematic language before presenting it to your learners.

3) Activities such as the ones suggested in this article may be time-consuming and allow little opportunity for communication. For this reason, they may work better as homework tasks. 


When even the most experienced language teachers look to their own knowledge for information about language in use, they will inevitably fail to provide their learners with the information that is important for acquiring it. It can be much better to look outwards, to language itself, and see how it is used. A corpus search provides the language learner with a convenient way of obtaining answers to linguistic questions.