Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
language is necessary and by what stage in order to tackle complex concepts such
as longitude and latitude in a geography context? Should learners be allowed to
believe that their performance is, on the whole, error-free? What happens to the
learners when the linguistic demands of the curriculum become very
sophisticated? How does a fluent EL2 speaker work out, without intervention, the
subtle but important semantic difference between to be frightened of and to be
frightened for in the analysis of the emotions of an army towards its general in a
history context? During one of my teaching sessions with a group of 14-year-olds
(one Chinese who was a fluent speaker of English and three EL1 students), none
of the students had understood the tasks instruction, requiring them to discuss
why an army was frightened for its general. All four students wrote about why the
soldiers were frightened of their general. Neither could they distinguish between
these two questions: (a) what do you think of the weather today? requiring
expressing an opinion; and (b) whats the weather like today? requiring a
statement of fact. The answer to both questions by all four students was: its
raining, which is the appropriate answer to (b) but not to (a). This example leads
us to ask WHY WE SHOULD CORRECT ERRORS - SLIDE 4
1.
self-repair
And this premise implies that learners must understand the feedback, and possess
the appropriate linguistic knowledge to self-repair (Karmiloff-Smith, 1985; Carroll
and Swain, 1993). Self-repair also depends on the types of errors that have been
produced (Corder, 1967), on implicit (Karmiloff-Smith, 1985) and explicit
linguistic knowledge (Krashen, 1985, 1987) of learners and on shared
metalanguage between learners and teachers (Carroll and Swain, 1993).
2.
transitional competence
Corder (1967) distinguished between two types of errors made by both L1
and Gombert, 1996; Donaldson, 1978; Downing, 1977; Titone, 1993) and to all
learning (Lee, 1997; Pratt et al., 1984). Pratt and Grieve (1984a) argued that as
children develop an awareness of the nature of cognitive functions, they gain
more control of thinking (Pratt and Grieve, 1984a, p. 9).
4.
fossilisation
HOW?
It is clear that corrective feedback should focus on Corders systematic errors
which may fossilise thus affecting adversely the development of a correct
interlanguage.
The timing of error correction
The timing of corrective feedback depends on whether it is a mistake or an error in
Corders (1967) sense. In the case of competence errors, the decision will be
dictated by a number of factors: is it a new error or one thats been encountered
before? Should it be dealt with immediately or should it be postponed (Allwright
and Bailey, 1991; Long, 1977)? There are advantages and disadvantages to any
timing. Immediate treatment interrupts learners flow and may not be positively
effective (Vigil and Oller, 1976). Postponing it to a future lesson will be less
effective, as time elapses between the error and the treatment (Chaudron, 1977,
Mina Drever, consultant, Training and Development Agencies for schools, London
minadrever@aol.com
1988; Long, 1977). Yet this may be necessary, particularly if the error is common
to the whole class (Holley and King, 1971).
Mina Drever, consultant, Training and Development Agencies for schools, London
minadrever@aol.com
Bohannon III et al., 1990; Brown and Bellugi, 1964; Brown and Hanlon, 1970;
Gordon, 1990; McNeil, 1966; Marcus, 1993). Some researchers saw expansions as
maternal imitations of the content of childrens utterances, albeit with added
grammatical forms, but they were not sure that this was feedback with any role in
language learning. Research in the last 10 years suggests that very young children
are capable of learning from some form or other of corrective feedback (Saxton,
1998, Saxton et al, 2005, Saxton et al., 2006).
Recasts have been extensively investigated in the observational L1 literature
but there is no agreement on their exact role in language acquisition (Bohannon III
and Stanowicz, 1988). Recasts are simple or complex (Baker and Nelson,
1984) reformulations of childrens utterances, which may or may not contain
errors. When they reformulate correct utterances recasts are modelling
structures that may or may not be already part of a childs current linguistic
knowledge (Farrar, 1992; Nelson, 1977; Nelson et al., 1973).
Feedback to children in L2
Studies in second language classrooms suggest that implicit corrective feedback in
communicative language teaching classrooms can have a direct effect on learners
output accuracy (Lightbown and Spada, 1990). Characterised by meaning
negotiation and meaning exchange leading to mutual understanding (Speidel,
1987, p. 104) they provide a great deal of rough tuning (Speidel, 1987, p. 110)
in the form of model/feedback, i.e. when teachers extend, expand, reformulate
recast learners responses, they simultaneously provide corrections and modelling
to learners linguistic errors.
However, Chaudron (1977) analysed teachers feedback in six of this type of
classroom talk (learners aged 1315) and discovered that SLIDE 7 - the most
effective corrective feedback first located errors and immediately rejected them, in
the form of repetition of the error with emphatic tone. Then metalinguistic
explanation for errors was provided to encourage learners to self-correct. In the
Mina Drever, consultant, Training and Development Agencies for schools, London
minadrever@aol.com
covert feedback was mentioned many more times, suggesting that covert feedback
was likely to be selected by more teachers than overt or mixed feedback. The latter
contained covert and overt feedback, making it a recipe for confusion and
ambiguity, as already found in the literature.
Covert correction to speaking was mentioned by an almost equal number of
questionnaire (19) and interview (16) teachers SLIDE 11 - . The predominant
mention of covert over overt feedback to speaking by both sets of teachers was
confirmed by the highly significant results of two Wilcoxon two-tailed signed
ranks tests. Among the questionnaire teachers, covert feedback to speaking was
mentioned by six times more teachers than those who gave overt feedback. And
more than three times the number of interview teachers said to give covert
feedback to speaking over those who did so overtly.
Summary of empirical results how teachers actually gave feedback
Eight video-recorded observation lessons were analysed to see if teachers taught
and corrected in reality as they claimed in the questionnaire. In-depth analysis of
teacher responses to errors provides further insights into teacher feedback. The
first thing SLIDE 12 to be said about the feedback given by the eight
observation teachers is that, overall, they corrected covertly with 76% of all
corrective feedback being covert. Secondly these 8 teachers corrected 71% of
childrens linguistic errors covertly SLIDE 13 while 18.4% went ignored,
because correction would break the flow of the interaction.
Interesting results from a finer analysis of teachers observed feedback
included the following: Corrective feedback was often confusing SLIDE 14 -.
Sometimes utterances in teachers corrective response wavered between overt and
covert negative, and I categorised it as covert/overt/negative feedback. At other
times it wavered between covert negative, implying rejection of pupils
Mina Drever, consultant, Training and Development Agencies for schools, London
minadrever@aol.com
10
productions, and covert positive, as if teachers could not make up their mind about
the appropriateness of pupils contributions. I categorised this feedback as
covert/positive/negative. Lyster (1998) also found ambiguity in the way teachers
used recasts, often accompanied by signs of approval when they responded
simultaneously to content and language errors. This is a problem specific to
communicative classrooms, suggested Lyster: - SLIDE 15 This reveals what must be a source of ambiguity for young L2
learners as well as a dilemma for teachers whose mandate is to
teach both language and content: namely, how to reinforce the
substantive content of student messages while giving them clear
messages about language form (Lyster, 1998 p. 71).
An example of a covert/negative/positive/feedback is this teachers sequence
of responses SLIDE 16 - . A group of pupils and teacher were discussing the
meaning of words with ch (pronounced as k) in them. At this point in the lesson it
was the word Christmas. The teacher had asked for a sentence with this word in it.
First she covertly approved (cvp) of it by repeating (rcr) a pupils correct response,
line 2. She overtly accepted it (ovp) in the next three utterances. Then she changed
her mind and she wanted more information (line 6) and asked a referential (rfl)
question. At this point the feedback reverted to covert, but I was no longer sure
whether it was covert/negative or covert/positive and categorised it as
covert/negative/positive (cnp). When she said that doesnt explain a lot (line 8)
she implied that the explanation given in line 1, which had already been accepted
as correct in lines 3, 4, 5, was now only partially correct. This confusion is
compounded by the use of tones 1 which denotes certainty in line 4 (very good)
and line 8 (that doesnt explain a lot); also tone 3 is neutral and could denote
certainty or uncertainty (lines 2 and 7, exact repetition of childs utterance).
Intonation contributed highly to all eight teachers idiosyncratic feedback. For
example one teacher gave exact repetitions in tone 5, which can express positive
Mina Drever, consultant, Training and Development Agencies for schools, London
minadrever@aol.com
11
12
13
14
Mina Drever, consultant, Training and Development Agencies for schools, London
minadrever@aol.com
15