(Part 1) Grammatical errors and an SR-EAL informed error analysis (A brief 'how to' guide!).

The first part of an essay that describes a project to put SR-EAL into practice by identifying the particular grammatical errors that appeared in my Thai students' written work.

Teaching English has its share of frustrating moments (for instructors and students alike) and we probably all agree that students' progress can often seem painfully slow. I've lost count of how many times I've walked into the staffroom to hear comments such as these from an exasperated and overwhelmed colleague:

"My students forget to provide the past simple tense after all that work!."

" EAL students struggle with subject/verb agreement!"

" EAL students forget to add the 's' plural marker to regular countable nouns -after all these years! "

And no doubt we could add a great many others to the list. If you've taught EAL for a term or two you come to accept that resolving such stubbornly persistent grammatical errors goes part and parcel with our daily job -indeed, for many of us it has become pretty much our whole job. We get through the day by telling ourselves that errors provide learning opportunities and are most certainly not something about which a child should feel ashamed or embarrassed. But that, of course, true as it is, must never deter us from exploring how to make our students' journey towards the 'error free' script just that little bit easier and, just as importantly, how we can remediate errors in the shortest instructional time. But what options do we really have at our disposal? What more can we do over and above what we offer already? To those who believe the answer lies in teaching the national curriculum or coursebook ever more diligently," I beg to differ: The gains from following that path depend upon how effectively your school teaches the curriculum already. And by and large international schools teach the NC very well already. I'd say you need to ask whether you can truly wring anything out of improving NC delivery yet further. Anyway, and I've asked this before, can we not surely do rather better than simply follow generic, one-size-fits all resources and ill-fitting curriculums adopted with little (if any regard) to a foreign setting?

So, how would an SR-EALer approach the challenge of improving children's written (or spoken) grammatical accuracy? In this, the first part of a three part blog, I define the term error, explain a system for categorizing these into meaningful types and, finally, briefly outline how I conducted an error analysis in a previous school (Thai children, primary). In Part 2 I look at the findings from the analysis, and in part 3 I suggest some activities that I have found useful for building children's grammatical knowledge. The essay has three aims: first, to give enough detail should you wish to carry out an error analysis in your own educational setting; second to demonstrate the power of error analyses to inform teaching in an international school setting and, third, to provide colleagues working in Thai based International education with useful data on error types and their relative frequencies (data on this appears in the next blog).

First steps -what is an error?

If we set out to equip children with the understandings they need to avoid 'serious' grammatical errors we need to identify those errors to give us the essential targets -the most persistent and intractable errors- around which to build future lessons; after all, we don't help students by teaching what they already know, unless we aim to consolidate prior learning. But this does raise the the critical question of what we mean by an error in the first place. Not all 'errors' we come across necessarily reveal deficits in underlying grammatical language competence. How many times you have asked a child to correct a sentence in their literacy book only to find they promptly do so -and without assistance from either peers or yourself. What does that tell us? Well... assuming the child hasn't learnt or acquired the relevant grammar since having written the text initially, then the error most likely falls under the label of pure carelessness -in other words, we do not have a true error at all, but rather what we should probably call a performance failing. The 'fix' for such 'slip ups' does not lie in grammar teaching at all (after all, the student has the requisite understanding); it lies in communicating to the student that s/he should check their work, and a reminder that in the future it would pay to focus rather more on the task at hand. Our remediation efforts would need to involve instilling a less casual attitude to writing. Perhaps we could spend time consulting works on student motivation and involvement seeking an answer there. The real concern -what we should our teaching efforts around- must remain the 'true' error: the error that arises from a deficit in the language user's underlying grammatical competence i.e. the basic knowledge (implicit or explicit) upon which s/he draws when communicating to others in either spoken or written form.

And that brings us to the issues of how to identify errors in a systematic manner, and the data source to consult with this aim in mind. For spoken English, a good source consists of all those events that oblige students to place a high regard on his/her grammatical accuracy. Formal speeches spring to mind, openings of class debates, or those occasions when you might have a student present an argument in front of the class. These are just the sort of occasions when speakers place a premium on grammatical accuracy and, as such, any errors we observe will likely reflect the true grammatical deficits we are looking for. Getting access to such data does present the occasional challenge, however. You'll need decent recording equipment and the technical skill to set it up to work correctly. You may even find yourself lumbered with the time-consuming task of having to transcribe recordings into textual form. If your interest lies, rather, in students written errors, data gathering typically proves a little less onerous. Exam scrips can prove useful, for example, since under 'formal testing' conditions children tend to put in that extra effort to supply error free work. If you work in a school that requires monthly (or half termly) unaided writing assignments you might consider using those scrips as the data source. As during exams, careless errors should occur infrequently since children will strive to produce their best work. The texts make for interesting reading and would seem a boon for anyone pursuing educational research into children's writing development.

How do I find and classify 'true' errors?

So, let's say you've gathered a random sample of monthly writing books for each of the year groups two to five. You now plan to use these as a basis for identify grammatical errors in the key stage 2 year group (though it could equally well be key stage 3 or 4; or even a single class). What do do you do next?

At this point I need to introduce the notion of obligatory occasion, a term I came across in a text by Burt and Dulay dating from the 1970s which explored learners' acquisition of grammatical morphemes. The authors define an obligatory occasion as a point in a sentence, or phrase, that requires a 'word' in a particular grammatical form.

Yesterday the teacher (look/looking/looked) at the homework.

Here we have an obligatory occasion for the past simple. Only the form 'looked' will ensure the grammatically correct 'Yesterday the teacher looked at the homework.' Had a student supplied anything other, we might assume the writer did not as yet understand the notion of past simple verb forms. Here's another example, this time for the definite article:

He looked up at (a/an/the) moon

A competent English speaker will tell you the correct choice is 'the.' The sentence does also illustrate the point that obligatory occasions can prove tricky to identify at times. Had I happened to pluck the sentence from a science fiction text set on a planet that happened to have three moons then we would likely have an obligatory occasion for the indefinite 'a'. Another example, again concerning articles, arises from the rule that we use the indefinite article when referring to something for the first time, and the definite article thereafter. The sentence:

A/The cat was walking slowly towards me.

Do we have an obligatory occasion for both 'a' or 'the?' As with the moons case (above) we would need to look at the surrounding sentences for clarification. For example, we could presume an obligatory occasions for the definite article if the preceding sentence read something like this:

I saw a large grey cat. A/The cat was walking slowly towards me.

The sentence below indicates an obligatory occasion for a comparative adjective form (faster).

He ran (faster) than all the others.

And this sentence:

He works hard everyday.

contains an embedded obligatory occasion for a present tense third person 's.'

The obligatory occasion gives us a systematic way of looking for errors, but still leaves the researcher with room for 'tweaking' the whole error categorization process. When carrying out any error analysis you soon learnt that you have to make important choices. Should I divide errors involving the articles (a/an/the) as simple 'article errors' or should I divide 'article' errors into their own distinct categories -e.g. failure to use 'a,' failure to use 'the,' and so on? The same problem arises with plurals. Do I treat a failure to add an 's' on regular plurals as one type of plurality error and failure to supply an irregular plural (mouse > mice) as a separate error of plural formation? To clump such errors into a single category conceals what amounts to two types of plurality knowledge and supplies rather less discriminating pedagogical data . On the other hand, this broad-brush approach makes error tabulation simpler and less time consuming. You will have to decide just how narrowly to define your error categories bearing in mind the time you have allowed yourself for completing the error analysis project and the learning goals you have for your students.

Furthermore, many errors will occur so infrequently that you have to ask whether you should categorize them under their own distinct headings at all, or rather simply assign them to a miscellaneous group. As a rule of thumb, about 12 error types account for the great bulk of errors you'll observe in written work. Errors that don't fall under any of your 12 headings might sensibly go into into the miscellaneous category; the choice depends upon the time you can spare for research and your pedagogical aims.

The actual analysis

Let's say, then, that you've collected up the children's' monthly writing books (or other source of data). Where do we go from here? Well, first off, you'll have to decide upon the type of study you have in mind. You might opt for, let's say, a longitudinal study in which you take samples of work for the same child from the time they attended year 1, to the time they finish year 6 (a within subjects experiment). You would end up with a record of the student's development over time, but not necessarily a picture of development among the larger population of which s/he is a member. Such a study could prove useful if, for example, you were working long-term with a single private student. Alternatively, you might take random samples of work from students in year groups of interest -i.e. samples from year 1, year 2 year 3 and so on (a between subjects design) and assume that the types of errors typify those of children within each of the year groups of interest. Whether you go for a 'between subjects' or 'within subjects' design comes down to where your research interests really lie. Neither design stands out as intrinsically more valid than the other. The choice really depends on where your research interests lie.

So, where do we go from here? With the notion of obligatory occasion to guide you, you can now move on to the all-important tallying stage. The real work now begins. You open a student's book, note the genre of the text you are about to analyze (I broke texts in to narrative, expository, descriptive and argumentation -the choice rests with yourself and other genres may recommend themselves) and then begin the error counting -placing each error into its particular category (past simple, article, 's' plural, possessive 's' etc). The whole exercise comes down to counting errors of each type. To get a feel of what this involves I reproduce a short text from an EAL Year 6 child writing a short biography of Roald Dahl.

Apart from the many examples of obligatory occasions, the text also also reveals the difficulties that error analysis can raise in practice. First off, you will need to decide just how detailed your error analysis need be. Even a script as short as this contains many error types. To identify all of these might prove a more challenging and time consuming task than you wish to undertake. Before carrying out your own error analysis it makes a lot of sense to look at a few scripts and note down those errors that seem particularly numerous (perhaps limit yourself to the most common 10-12 as I suggested earlier). These will then make up the particular errors you go on to categorize. As for other errors -those you assign to a miscellaneous category (see above)- you could always revisit and categorize them at a later date if so inclined. If you do insist on categorizing all errors, bear in mind (as we will see in Part 2) that a relatively small number of error types comprise a very large proportion of the total errors you observe in any one script. This emerges clearly from my own error analyses (I've conducted 3 so far, one of which I present in Part 2) and likely holds just as true for children attending your school, too. Believe me, you really don't want to attempt the mother of all error analyses only for the task to become disheartening and unmanageable further down the road.

Note also, that some sentences will contain two obligatory occasions conflated into one and you will need a systematic way of handling these. Take, for example:

He play tennis yesterday

We have an obligatory occasion for the past simple (played), but also an indication that the writer does not understand subject verb agreement. My instincts tell me that I should record an instance of a past simple failure, and also a subject verb agreement error.

A final point concerns students' L1. Bear in mind that the types of errors you observe in a script (or recording if your focus is spoken English) will typically relate to the child's mother tongue. you might sensibly decide to note the particular errors for each of the various L1 populations you work with. A child form France, for example, will tend to make rather different grammatical errors to a child from Thailand. If you group children together irrespective of mother tongue you run the serious risk of missing common errors associated with particular student language backgrounds. As we will see, in Part 2, my own error analysis focused on Thai L1 EAL learners for this reason -not that my learners from France or Russia deserved less attention but because it made sense to initially address the numerically largest population first and it so happened that most of the class consisted of Thai mother tongue speakers.

Conclusion

As noted in the introduction, if we're serious about raising grammatical correctness, then we have to set out to identify those errors we deem important in the first place. Once we find those errors, get an idea of their relative proportions and just how numerous they are, we can then then go on to think about designing a bespoke instructional program with resolving those particular errors foremost in mind; such is the SR-EAL way! Burt and Dulay's notion of obligatory occasions offers a systematic and principled way of identifying errors in both written and spoken English and I hope I have managed to provide a few pointers should you take up the challenge of conducting your own error analysis research. This essay only covers basic aspects of what is error analysis, a topic on which you'll find no shortage of excellent and informative texts and 'how to' guides; for a good introduction, you might try Pit Corder's Error analysis and interlanguage (1980) or perhaps James (1994). In part 2 I go on to look at the results of a typical error analysis study.

References.

Corder, S.P. (1967). The significance of learners’ errors. International review of Applied Linguistics, 5, pp.161-170 and reprinted in J.C. Richards (Ed.) (1974). Error Analysis: Perspectives on Second Language Acquisition. London: Longman.

Corder, S.P. (1971). Idiosyncratic dialects and error analysis: International review of applied linguistics, 9,(2), pp.147-160. Corder, S.P. (1972). Describing the language learner’s language: In CILT Reports and Papers, 6, pp.57-64.

Corder, S. P. (1981). Error analysis and interlanguage. Oxford: Oxford University Press

James, C. (1994). Don’t shoot my dodo: On the resilience of contrastive and error analysis. International Review of Applied Linguistics, 32,(3), pp.179-200.