Professional Documents
Culture Documents
Analysis and
Correction
Summary
In this input we will be looking at the issues of error analysis and correction. We will review
attitudes to error within different approaches to language teaching and learning. We will be
considering why learners make errors and the types of errors they make and then we will go
on to consider possible teacher responses to error. Are mistakes ‘bad’? Do your learners
want all their mistakes to be corrected? What methods can we use for correction? Is
correction useful? You will be asked to consider all of these questions within this section.
We will also summarise some of the more recent, and controversial, attitudes towards
learner error.
Objectives
By the end of this input you should:
Contents
1. Introduction
5. Categorising Errors
8. Terminology Review
Reading
1. Introduction
Think of something that you were taught to do: this could be something learnt in a formal
educational environment or a practical / vocational skill. In that learning situation, both you
and your teacher would have had attitudes to the mistakes you made and your teacher
would approach your errors in a certain way. For example, they might have identified what
you needed to improve, or they may have just highlighted your faults. Whatever their
approach to correction, it may or may not have met your preference but it would have
certainly made you feel in a certain way. Perhaps more than anything, our view of error and
how it is dealt with determines how we view a learning experience. Whilst a memorable
teacher may be one who, by whatever means, helps us to improve our performance in a
given area, it is also possible that how you were corrected is crucial in determining whether
you remember that learning experience as positive or negative.
The most efficient materials are those that are based upon a scientific description
of the language to be learned, carefully compared with a parallel description of the
native language of the learner.
So, in practice, CA consists of the comparison and contrast of the linguistic systems of two
languages, for example the grammatical or phonological systems of French and English. The
basic premise of CA is that the main difficulties in learning a new language stem from
interference caused by the differences between the target language and the learner’s first
language. So, we might assume that the use of the auxiliary in the French past tense e.g. ‘J’ai
vu ce film hier’ will cause a French learner of English to produce the error I have seen this
film yesterday. Proponents of CA held that, through contrastive analysis of the learner’s own
language and the target language, main problem areas could be identified and teachers
could plan to teach key differences in order to lessen interference.
The rejection of behaviourism and CA, following the work of Chomsky from the 1960s
onwards, was reflected in the cognitive approach’s attitudes to error and correction. In
theory, teachers became more aware that errors are an inevitable part of learning, that they
provide evidence of learning taking place and that many of them will right themselves in
time. For example, the overgeneralisation of rules which present themselves as error e.g.
the use of an –ed ending for all past tense forms could be viewed as the learner applying a
newly-learned rule analytically to generate new language, albeit incorrectly in certain cases.
In this way, errors can be seen as evidence that learning is taking place, and can also serve as
a way of gaining insights into the language acquisition process.
As language learning came to be seen as a cognitive activity, teachers were encouraged not
to correct mistakes so much as simply to indicate that they had occurred, encouraging the
learners to reflect on and work out the error, and to arrive at the ‘correct’ version through
the conscious application of language rules.
As the distinction between ‘learning’ and ‘acquiring’ language began to take hold by the late
1970s, errors came to be viewed as a natural aspect of the learner’s development. The best-
known commentator on the distinction between learning and acquisition, Krashen,
suggested that interventions by the teacher in order to correct errors are useless or even a
hindrance when the students are focussed on getting a message across. He says:
Error correction has the immediate effect of putting the students on the defensive.
It encourages a strategy in which the student will try to avoid mistakes, avoid
difficult constructions, focus less on meaning and more on form. It may disrupt the
entire communicative focus on an exchange. […] Since error correction is not of
direct benefit to language acquisition, […] a safe procedure is simply to eliminate
error correction entirely in communicative-type activities.
He proposed that correction when focussing on fluency is, at best, limited in its effect
because learners proceed through a ‘natural order’ in language learning. In other words, we
try to correct areas of language which students are not yet ready to produce accurately.
Certain errors are bound to be repeated because they are introduced in the language
classroom at too early a stage in the language learner’s development. Note that Krashen
does permit the correction of errors when students are attending to form, however,
although he makes it clear that this is not a magic wand which will instantly result in
accurate language use.
As a more communicative approach to language learning and teaching took hold, stemming
from Krashen, proponents of a ‘strong’ communicative approach argued that learning only
takes place during interaction with a communicative objective. Within this context,
grammatical error took a back seat, with the main aim of teaching and learning being the
successful negotiation of communication. Problems of error were to be resolved only when
they led to communication breakdown.
From this wide range of theories about how language is learned we have garnered a
corresponding range of attitudes to error. For many teachers (and perhaps more learners)
‘correction’ is still high on the agenda of classroom activity. Whilst many learners would like
all their errors to be addressed, there are many teachers who feel that only errors which
interfere with communication need to be dealt with. Errors and how they are dealt with are
subjects which can divide staffrooms, with some teachers taking steps to prevent errors
becoming ‘entrenched’ through strict approaches to correction, and others viewing the
communicative goal as overriding the goal of linguistic accuracy. However, wherever you
stand on this correction continuum, you need to consider carefully the reasons why learners
make particular errors and the effects that their attitudes to error and correction might have
on your own teaching.
The view of the causes of error mutated from one where the mother tongue was seen as the
source of error to one where error was seen as a natural part of language learning. This view
of error as evidence of a cognitive process found expression in Pit-Corder’s scheme for the
classification of error in Error Analysis and Interlanguage, Oxford University Press, 1981. He
initially suggested that there are three types of error:
Systematic error (the learner has found a rule but is applying it wrongly)
Post-systematic (the learner has lapses in his/her use of the correct rule)
However, he later revised this view to simply pre- and post-systematic errors, feeling that
systematic and post-systematic errors both amount to the same thing.
Pre-systematic error
At this stage the learner is not aware of any rule at all, or is confused about the rule, so it is
not yet part of his/her system for using the language. Pre-systematic errors often result in
learners ‘borrowing’ a rule from L1 or using an inappropriate rule from the target language.
At this stage, errors in the use of an item will tend to be random guesses with no system. In
the same text the learner may produce:
Clearly the learner is not at all sure how to mark the 3rd person present simple and may be
trying each form out to see which is right. The more complex the discrimination the learner
needs, the greater the number of wrong guesses. Learning the linguistic forms of the present
simple is in fact quite complex because of the numerous elements involved when
formulating it.
Post-systematic error
At this stage, the learner is either still constantly making errors but each of the errors has a
definite pattern. They have discovered a rule, but not the rule, and are applying it
consistently, meaning that they will be consistently wrong. In contrast, in the pre-systematic
stage they could chance upon the right rule or have internalised the correct rule but use is
still rather conscious and errors will probably occur in less controlled contexts where the
need for communication overrides the need for accuracy. The latter are usually the kind of
errors which can be corrected by the learner when attention is drawn to them.
For a very succinct and readable account of Pit-Corder’s classification of errors, see Ellis, R.
1997 Second Language Acquisition Oxford University Press, pp 93 - 94 (Text 3)
Pit-Corder’s classification system provides teachers with a useful tool for analysing, and
more importantly, prioritising learner errors. By recognising that learners can make errors
because they are being adventurous i.e. that they are taking risks with language which take
them beyond the boundaries of their knowledge of that language, this system of
classification enabled teachers to diagnose areas which might need to be covered in addition
to those which needed review.
Using Pit-Corder’s scheme, we can assume that learners will make pre-systematic errors
when expressing themselves above their level, when the teacher has been unclear or
overloaded them, or when for some reason they cannot see a rule.
Post-systematic errors will occur as a natural part of the learning of a new item, and show
the teacher what angle to take in remedial work or further practice activities. They will recur
as the learners move on to learn new items, thereby shifting their focus on to new things
and concentrating less on the old. Basically, post-systematic errors are ‘slips’, which are
made when people are tired, or under pressure.
(from Ellis, R. 1997 Second Language Acquisition Oxford University Press, p16)
It is clear that the learner has made a number of erroneous forms here. One way to identify
errors is to compare them with that which is considered ‘correct or ‘standard’ by native
speakers. Bearing this in mind, then, A man and a little boy was watching him should
therefore be corrected to A man and a little boy were watching him. But what about, for
example, The big of them contained a snake? Would a native speaker naturally render this
as The bigger of them contained a snake, or The big one contained a snake? Identifying
exactly what students are having difficulty with is not always easy. And even comparing
erroneous forms with what we feel to be more natural is problematic for what exactly
constitutes a ‘correct’ utterance? went in the traffic would probably be more native-speaker
like as they went into the traffic, but in is perfectly grammatical, and possibly what native
speakers produce when hurried or stressed. Finally, let us consider these two samples: The
basket contain a big snake and The big of them contained a snake. The former would
indicate that J has problems forming past simple regular forms properly. Yet earlier he has
produced it accurately. It is clear that J is aware of the rules which govern this item of
language, yet at times does not adhere to it. Perhaps by the end of the sample he was tired,
or perhaps attending more to the lexical item basket (quite a low-frequency item), or
perhaps the rule is not fully concretised in his brain. For whatever reason, though, we cannot
say that this is an area that J needs to be taught. It is because of this kind of evidence that
we distinguish between errors i.e. those erroneous samples which occur, like Pit-Corder’s
pre-systematic errors, before the language has been accurately internalised, and mistakes /
slips, which occur after the language has been attended to, but which come out incorrectly
from time to time. As native speakers, we know what we ought to produce, so we cannot
really be described as producing errors in our L1, but we do make dozens of slips daily. While
authors often use the words mistake and slip synonymously, it is worth reflecting that we
would have difficulty describing native speaker slips with spelling and grammar as mistakes
unless there were a number of inconsistent language samples with the same or similar
problems. With learners also, it is worth considering the frequency and impact of faulty
language samples before choosing to use the term mistake or slip.
5. Categorising Errors
As the sections above demonstrate, research into language learning has identified a range of
reasons why learners might make errors. Just as current teaching techniques draw on a
range of approaches which developed from opposing camps within the field of
psycholinguistics, so too we can identify ways in which the evolving theories about the
causes of error still influence current thinking on the subject.
Categorising errors can help the teacher to understand why the learner has made the error
and can help the teacher to formulate a strategy of error treatment. The following
categories are a summary of Bolitho and Tomlinson’s taxonomy in Discover English (pp 114 -
115):
L1 transfer
False analogy
The learner has compared a language item and made an untrue comparison e.g. ‘plait’
rhymes with ‘wait’.
Overgeneralisation
The learner applies rules too generally e.g. the use of ‘will’ for future meaning.
Overlearning
The learner may have become too focused on a piece of language because it has been
taught recently or because it does not occur in his or her own language and
subsequently overuse it.
Ignorance
Incomplete learning
The learner has not learnt the whole rule, only part of it.
From your own experience, you may like to add to these categories. For example, learners
may make performance mistakes because they are tired or hungry, or because of the
teacher’s own (lack of) competence, or by being overly-obsessed with accuracy.
Of course, identifying and classifying error provides useful data for teachers, but does not
help learners eradicate problem areas. We need to think about how best to correct. We will
now move on to consider how to approach error in the classroom.
However, the blanket approach described above is one which is often demanded by
learners. It is likely that learners’ expectation of error and error treatment will be based on
their previous experience. Consequently there are learners who feel that, if the teacher does
not pursue a policy of strict correction of every error, s/he is not doing their job properly.
How can this be addressed? Some teachers find that it is a good idea to discuss their own
policy of error correction with their learners at the start of a course and review this, again
with learners, at later stages in the course. Encouraging learners to see making errors as an
essential and positive part of the learning process may take time but, if learners are to
develop skills to enable themselves to learn independently, the acceptance of error is
essential. However we choose to deal with error in the classroom, it is important that both
we and our students understand the approach being taken, and its implications. Although a
teacher may feel that, in communicative activities, direct correction might interrupt the flow
of conversation, students within the activity might feel that correction is essential to their
success in the activity. Similarly, when a student asks a teacher to correct every mistake they
make they may not realise that such an approach could limit the time spent on new items,
or, if adhered to closely, could undermine their confidence in their own abilities because of
the number of errors being made. It is vitally important to make students aware of our
overarching policy towards correction, and also how correction will be managed at different
stages of the lesson.
1. ‘I tend to correct at all stages in the lesson. My students are very accuracy-focused.’
3. ‘The amount I correct depends on the students’ mood and my own. I don’t have any firm
policy.’
Obviously, the reaction to all of these statements is ‘It depends’. However, here are the
views of a Delta tutor interviewed on the topic.
1. ‘Some students, and not only accuracy-focused students, respond well to correction at
any stage within a lesson. For example, students who are confident and fluent might find
it useful to have on-the-spot feedback on their errors, regardless of the aim of the
activity. This approach may be more common at higher levels higher levels, where
focusing on dealing with error is one way of highlighting to learners the areas which
need attention.’
2. ‘I tend not to interrupt the students’ flow of thought, but I’m not averse to providing
correction during an extended activity. Although time spent on errors at the end of an
activity can provide focused attention on general areas of need, sometimes learners are
unaware that they have made a particular error so saving it till later can reduce the
impact of the correction. On a purely practical note, it’s also very easy to run out of time
for dealing with errors if everything is kept until the end. If you are taking this approach,
remember to plan in ‘error’ time!’
3. ‘Yes, I agree with this one. As a language learner, I know that there are days when I’m
really eager to get everything right and others when I’m really just too tired. It’s
important to gauge the mood of learners when approaching error and feedback. Why
not ask your groups from time to time to choose which approach they would prefer on a
given day?’
Few of us take a completely blanket approach to error correction, preferring instead to focus
on certain types of error, perhaps persistent errors such as a group of advanced students
omitting the 3rd person ‘s’, or varying our approach depending on the aim of the classroom
activity. Whatever influences our choice, the selection needs to be conscious in order that
the approach to error, and the way in which it is dealt with, develops as a coherent policy
rather than the ‘scattergun’ effect which we have when a wide range of discrete errors are
dealt with in one go. It is perhaps useful to discuss with your students what types of errors
they feel should be dealt with in the classroom. In this way, the error-focus activities take on
the features of learning activities within themselves, with learners being better able to
recognise the value of the task.
The aim of a classroom activity is the predominant deciding factor in choosing when to deal
with error. You may be a teacher who insists on accuracy in controlled practice tasks, acting
to correct promptly, but you may allow error to pass in fluency tasks. You might be of the
opinion, however, that discreet correction during fluency activities is effective because the
student becomes aware of the error immediately and has a better chance of being able to
self-correct at a later point. Or you may ‘save’ errors for later follow-up work. Whichever
approach or combination of approaches you choose, you need to be aware of why you are
making that choice. Consider what underpins your decision to act immediately, to save
errors for later, or to let things pass. The factors might include the students’ attitudes to
correction, their learning goal for that particular lesson, whether the work you are doing is
an initial introduction to a language area or remedial work or how your students are feeling
on a particular day.
You will have noticed that for the most part we have used the phrase ‘deal with error’ rather
than the verb ‘correct’ in this section. This is because teacher correction is only one of a
range of options we have at our disposal when faced with error. Other possibilities include
self and peer correction, where the student who has made the error or a classmate provides
the correction, and diagnostic work where the errors go uncorrected but provide the
teacher with a clear indication of areas which need to be included in the teaching
programme.
Finally, in place of correction you might consider reformulation of the language which the
students produce. Correction tends to focus on the ‘putting right’ of discrete items which
the student has produced incorrectly. Reformulation, on the other hand, focuses on longer
chunks of language which may or may not contain grammatical, lexical or phonological
errors but which could be limited in terms of style. The aim of reformulation is to ‘upgrade’
the student’s language e.g. to provide a more idiomatic expression where a formal chunk is
used: ‘Is the film worth seeing?’ as opposed to ‘Do you recommend me to see the film?’
Reformulating the student’s use of language allows the teacher to focus on items which,
whilst not inaccurate, do not reflect an idiomatic use of English.
The level of the learner. We need to recognise if the error is the result of a low-level
student experimenting with new language or a higher level student making a slip.
The focus of the activity. Was the activity designed to give practice of specific language
items? If so, errors within those items might be seen as ‘important’.
The first two criteria focus very much on the place of error in a communicative context. At
the point where communication breaks down or is adversely affected by the speaker or
writer’s errors, some type of intervention is necessary to help ‘repair’ the communication.
Errors of this type are referred to as ‘global’, and are often contrasted with ‘local’ errors, i.e.
those which may sound unnatural but do not cause a breakdown in communication.
The third criterion is focused on the individual student and the extent to which this error is
representative of the type of errors made by the class as a whole. As teachers, our decision
is whether intervention through some form of correction would benefit this particular
student or if the error indicates that the student or group as a whole are now ready for the
introduction of new language items.
The fourth criterion mentioned above addresses the learning aim for the class as a whole.
This is where our considerations of whether an activity has an accuracy or fluency focus
might dictate how much correction we do, and the form which that correction might take.
For example, we might adopt a higher corrective stance during accuracy activities, where the
focus of the exercise is on the form of the language in any or all of the language categories.
On the other hand, we may take on a generally non-corrective stance during fluency
activities, where the focus of the exercise is on effective communication and the
achievement of a task-based objective.
Selectivity is another guiding principle. We need to choose which types of error we are going
to focus on in any particular activity; otherwise there will be simply too much to correct. In
an accuracy activity, correction emphasis may be purely on getting the syntax and rhythm
right, or on correct production of sounds, or intonation. In a fluency activity, if you are taking
a generally non-corrective stance, you may decide only to intervene in cases where learners
are unable to make themselves understood. You may decide to take note of errors which
occur in group or class work, which centre on general areas of weakness, or how best to
organise the information load, and then feed these back to the learners afterwards for
comment and correction.
In a semi-controlled activity, where there may be revision and practice of a wider range of
language items, correction emphasis may be on general rhythm and intonation, and on lexis,
or on other errors, which are frequent among the particular group of learners.
The important things to bear in mind always are the priority considerations. Ask yourself the
following questions: is it a pre-systematic or post-systematic error? Is it a word or a structure
which has recently been taught and is therefore still in the process of being assimilated? Is it
global or local? Is it a frequent error or an infrequent error? Is the activity controlled or free?
And so on.
If the learner cannot correct their own error it is often useful for other learners to become
involved in the process through peer correction. In this way the group are being asked to
review their knowledge and contribute in a supportive way, reinforcing language learned.
Approached in the right way, peer correction can help to foster a mutually supportive
environment where learners can help each other and reduce their reliance on the teacher as
the sole source of knowledge. If learners cannot help each other it is a useful sign to the
teacher that some remedial work is in order. However, it needs to be handled delicately. If
the same student always provides the correction, this can be divisive for the class. Some
learners might not like to be corrected by peers. There may also be socio-cultural issues of
age, status, and sex involved.
There is a variety of techniques which the teacher can employ to give immediate feedback
on spoken errors. Here are a few:
Questioning facial expressions can be used to show that an error has occurred.
Hand gestures can highlight specific errors such as tense, contractions, word stress,
word order and missing words.
Overt correction can be used to tell learners that they are wrong e.g. ‘No. That's not
right. It's…’
Reformulation
Echoing, where you repeat the error the learner has made, but with a questioning
intonation.
All of the techniques above presuppose that you have made the decision to give ‘on the
spot’ feedback. As we discussed before, we may not always want to interrupt activities to
correct and may decide on a policy of delayed correction.
Take notes of learners’ mistakes and write them on the white board after the activity.
Invite learners in pairs or groups to suggest corrections or come to the board and write
their suggestions. If you have access to an OHP you could save time by writing directly
on to a transparency which could be photocopied later. It is useful to put mistakes into
categories e.g. pronunciation, grammar, or vocabulary, as this helps to focus learners
and avoid guessing.
Hold a grammar auction. Have a list of correct and incorrect sentences produced by the
learners in a previous lesson (or lessons) and ask them in pairs or small groups to decide
if these sentences have an error in them. Give each pair or group a notional amount of
money. Offer each sentence for auction. The learners should try to buy the correct
sentences. The pair or group with the most correct sentences and most money left wins.
(From Rinvolucri, M. Grammar Games Cambridge University Press).
For more personalised correction take notes of individual learner's errors and provide
class time for each student to analyse and correct his/her own errors. This may not be
too difficult with a small class but you may choose to focus on only pairs or groups of
learners with larger classes.
Responding to output in this way is vitally important in helping learners develop and hone
their linguistic (and communicative) competence. Picking up on salient elements of their
output and focussing on them is often referred to as dealing with emerging (or emergent)
language. It often takes the guise of error correction, but we can also pick up good elements
of the learners’ output and share them with the whole class, and feed in more natural and
idiomatic language through formulation.
Written errors
As with speaking we have the options of teacher, self and peer correction available to us.
Again we have to decide what we are going to correct, and it is useful for the learner to have
errors highlighted and categorised if they are to attempt self or peer correction. For this
purpose we might decide to use a correction code or list of symbols to categorise mistakes.
Here is an example of a piece of student work where the teacher has used a correction
code:
Correction code
S- spelling G -grammar
The grid below outlines a range of practical activities and an evaluation of each technique.
Gesture from Does not interrupt the flow of You need to train learners to
teacher the conversation. recognise the gestures you use.
Self Correction Very useful for slips and The learner may not be able to
systematic and post-systematic self-correct as they do not
errors. know the answer.
Echoing by Highlights the fact an error has The learner may not recognise
teacher been made. this as correction and not
attempt to correct the error.
Reformulation by The learner is given the correct The learner may not recognise
teacher form in a sensitive way that this as correction and ignore it.
native speakers themselves use.
Grammar Good fun and learners get a Not personalised, some learners
Auction chance to see spoken errors may not have any of their own
written down and have more errors included.
time to try and correct them.
White board Gives learners more time to Again learners may not be
/OHT reflect on their errors and help focusing on their own problems
each other. but on other people’s.
8. Terminology Review
For each of the terms below, think of a definition.
Example:
Error analysis: This is the field of applied linguistics which collects and assesses learners’
errors in order to analyse what learners get wrong, why, and what we need to do in
classroom in order to minimise those errors.
1. Global error
2. Pre-systematic error
3. Mistake / slip
4. Post-systematic error
Suggested Answers
1. A global error is one which interferes with the comprehensibility of an utterance.
2. Pre-systematic errors occur prior to the learner’s being exposed to the language item in
questions.
4. Post-systematic error is an error which occurs after the student has been explicitly
exposed to the rule.
Reading
Although not essential to your Module 1 preparation, if you would like to explore this area
further we suggest the following:
Suggested Reading
Bartram, M. & Walton, R. 1991 Correction LTP (Chapters 3, 5 & 7)
Ellis, R. 1997 Second Language Acquisition Oxford University Press, Chapters 2 & 3
Ellis, R. & Barkhuisen, G. 2009 Analysing Learner Language Oxford University Press
Additional Reading
Bolitho, R. & Tomlinson, B. 1995 Discover English Heinemann Unit 6
Johnson, K. 1999 Forty-two teachers, one learner: correcting written work MET 8/4
McPherson, K. 1995 When and how to correct or not to correct? MET 4/1
Parrott, M. 1993 Tasks for Language Teachers Cambridge University Press (pp70-71)