Elgazette Logo newtrans  The magazine for English language teaching and English medium education

Interesting facts about grammar (really)

Anne O’Keeffe and Geraldine Mark discover what students really learn – and when

We have spent the last four years researching learner grammar using the Cambridge Learner Corpus and have found many surprises. We were asked by Cambridge University Press and Cambridge English (as part of the English Profile project) to come up with a list of grammar items that learners know – or ‘can do’ at each CEFR level. By this we’re talking about what they can do with forms, meanings and uses.


You might ask, ‘Why not just look at the coursebooks and the syllabi?’ Yes, but we wanted to know what grammar learners actually know based on what they write in their exams, not what grammar we think they know because we taught it!

So what we have come up with is English Grammar Profile. It’s a list of 1,222 grammar items that learners can do at each of the six CEFR levels, under nineteen super-categories. In order to arrive at these, we first had to make a list of all the grammar that is typically taught in ELT by looking at coursebooks and grammars. This gave us the super-categories (such as prepositions, passives, reported speech and future). For each of these, we combed through the 55 million-word Cambridge Learner Corpus. This corpus contains over 200,000 exam scripts from 215 countries by speakers of over 143 first languages, and 32 million words of the corpus have been coded for grammatical errors. Here are just six of the many insights from our research.

We are delighted that CUP has made the English Grammar Profile available free online as a resource. We really hope that it will help us rethink what grammar we teach and when we teach it based on what learners really do with language. Check it out yourself at www.englishprofile.org/english-grammar-profile.

Grammar can be right but wrong

When we investigated a grammar item, we always started at A1 and A2 levels to see what was happening at the lower levels, and often we were surprised to find evidence of correct use of grammar, though in a limited way, at these levels. However, sometimes we would find correct grammar patterns which were used in the wrong way. The best example of this is in relation to modal verbs. We found all of these modal verb examples used by A1 and A2 learners, and while the grammatical pattern is correct (pronoun + modal + main verb), you would really lose your friends if you used modals like this in the real world as they are pragmatically incorrect because they are just too direct:

You will attend my house at 7 o’clock.

Hello Sam, you have to come at my home at 7.30. You should bring a ball.

So getting grammar right doesn’t always mean it works in every situation, and this is something learners pick up as they go up the levels.

Conditionals are learnt before you have finished teaching them

Sometimes we were surprised to see that learners could do things earlier than we would have expected. The best example of this is that all of the conditional clause patterns that we typically teach, and more, were competently used by learners by B1:

A2: If + Present Simple. If you buy a gift for me, please buy a book, because I love to read.

If + Present Simple, with an imperative in the main clause: If you want to come, just tell me.

At A2, they can use the if clause to soften the directness of a suggestion or offer:

If you like, you could bring some drinks.

By B1, learners can do the following with if clauses:

First conditional: If you live in the countryside you will see beautiful nature.

Second conditional: Maybe it would be more fun if you went with your friends.

Third conditional: If they had not been there, I would have been very lonely.

And they can use conditional clauses in other ways that we wouldn’t expect at B1, for example to talk about an exception, with present and future reference:

If I were you I’d go to work on foot, unless it’s far away from your house.

There’s lots more to say about conditionals. It may be that the EGP will force us to revise grammar syllabi to align properly with what learners can do based on this kind of real evidence and what native speakers of English actually do with conditionals.

Sometimes the grammar we teach isn’t all that commonly used

Another eye-opener for us was when we found that some grammar items were used more by learners in exams than by ordinary people in their everyday use of English. The best example of this is the dreaded future perfect continuous. How often have you struggled to teach structures like ‘By ten o’clock, I will have been working for two hours’? Think about it. How often do you actually say things like that? When we had a look at the data in the Cambridge Learner Corpus for the future perfect continuous and compared it with the British National Corpus (a corpus of written and spoken British English), we actually found that our learners use it more frequently than we do in the real world! Here are some learner examples from the CLC:

‘This summer I will have been working for three years for my company.’

‘Now I am staying in Hastings in England and this month, I will have been studying English for seven months.’

When we compared how frequently the present perfect continuous was used in the Cambridge Learner Corpus with how often it is used in the British National Corpus, we found the frequencies to be low in both cases but overall lower in the native speaker corpus (0.2 per million words compared to 0.3).

You can’t count on countability even at C level

Students at C2 level are still having problems with these even though the grammar point is taught in most coursebooks at A1 and A2 level. Looking at the corpus, these are the most error-prone uncountable nouns: informations, advices, equipments, etc.

These words are not ones that A1 or A2 learners would know, so the important point is that we teach learners that some nouns are countable and some uncountable at A1 and A2 when they only know a small number of nouns. As learners advance their levels of English, they learn lots more nouns and the risk of making a countability error is much higher. We need to keep recycling our pedagogical focus on countability right up to C2 level.

Pronouns are still a problem at C level


One of our big surprises was to see how some seemingly basic grammar items, which are taught to learners at an early stage, were problematic even at C levels. Take pronouns. According to the CEFR, demonstrative, personal, subject and non-subject and possessive pronouns are all listed as can-do statements for grammar at A2 level. When we looked at the learner corpus, it wasn’t as neat and tidy as this. By A2, learners did know many of the pronouns but they still had a lot of problems with others. Just looking at the possessive pronoun theirs, for example, we found that, compared to the other possessive pronouns, it was problematic right up to C level. Here is a graph showing how learners are still getting it wrong at C1 level.