Teachers research teacher

talk
Scott Thornbury

This article describes an in-service training project designed to raise

trainees' awareness, through the analysis of transcriptions of teaching
sequences, of the degree of communicativeness in their classroom
interactions. The presence or absence of such features of communicative
classroom talk as referential questions, feedback on content, wait time,
and learner-initiated interaction, are used as 'bottom-up' markers of
communicative, content-driven, teacher-student interaction. Trainees'
analyses showed evidence of growing awareness of their non-
communicative ritualized teaching behaviours, awareness that, at least in
some cases, resulted in improved classroom practice.

Introduction It has become something of a truism to suggest that self-styled

'communicative' teachers are seldom, if ever, communicative in practice.
Using transcripts of classroom interactions, Nunan demonstrated nearly
a decade ago that 'there is growing evidence that, in communicative
classes, interactions may, in fact, not be very communicative after all'
(Nunan, 1987: 144). This claim was echoed by Kumaravadivelu, also
citing lesson transcripts: 'Even teachers who are committed to CLT can
fail to create opportunities for genuine interaction in their classrooms'
(Kumaravadivelu 1993: 13). Legutke and Thomas (1991: 8-9) are less
circumspect:
In spite of trendy jargon in textbooks and teachers' manuals, very little is
actually communicated in the L2 classroom. The way it is structured
. . does not seem to stimulate the wish of learners to say something, nor
. does it tap what they might have to say . . . Learners do not find room to
. .• • speak as themselves, t o use language in communicative encounters, to
: create text, t o stimulate responses from fellow learners, or t o find
solutions to relevant problems.

In-service training Unhappily, the above description accurately reflects the kinds of classes
that are frequently observed on in-service training courses, such as those
preparing for the RSA/Cambridge Diploma in TEFLA. If challenged,
the teachers generally claim allegiance to a communicative methodol-
ogy, adducing as evidence the essentially interactive nature of the
lessons. But, while there is often considerable interaction between the
learners in these classes, for the most part such canonical techniques as
information-gap activities rarely do more than mediate exchanges of
meaning at the most trivial level. Since there is seldom any real incentive
to communicate, discernible engagement on the part of the learners is
ELT Journal Volume 50/4 October 1996 © Oxford University Press 1996 279
 minimal. Likewise, while group work is a commonly used participation
structure, it is frequently conducted at the lowest possible level of
involvement; there is often the feeling that the only real communication
between the learners takes place during the break, or after class.

Eliciting In these lessons there are often prolonged sequences of teacher-fronted

eliciting of the classic IRF type (teacher initiates -> student responds ->
teacher follows up/gives feedback):
[I] T: Um, a little bit of vocabulary. Let's look at the picture. What
can you see?
[R] SI: Children.
[F] T: You can see some children. You can see a TV.
[I] What else can you see?
[R] S2: A sofa.
[F] T: A sofa.
[I] Or another word, same thing, different word.
[R] S3: Couch?
[F] T: Ah! Couch. A couch.
[I] Everybody: couch.
[R] Ss: Couch.
[F] T: Couch. OK?
(Diploma trainee's data)
While such sequences often give the illusion that classes are interactive,
and that the learners are equal participants in the 'conversation', very
little here is qualitatively different from the question-and-answer
routines that typified methods thought to be much less 'communicative'.
For example:
Teacher Pupils
What's this? It's a stone.
Is that the table or the floor? It's the table.
Is that the window? No, it isn't.
Is this a match? No, it isn't.
What's that? It's the wall,
etc. etc.
(Palmer 1940, 1970: 23)
This basic mismatch between principle and practice has been attributed,
among other things, to the constraints imposed by the grammatical
syllabus, which, as Willis (1990) convincingly argues, is incompatible
with a truly communicative methodology: where grammar is the agenda,
whether overt or covert, then grammar rules. This grammar-driven
conception of teaching is exacerbated by training programmes which
place a high premium on achieving lesson aims. These aims are usually
couched in terms of discrete pre-selected linguistic items, and their so-
called 'achievement' is often at the expense of spontaneous, student-
280 Scott Thornbury
 generated interaction. Teaching materials may also be partly to blame.
As Swan (1990: 94) observes: 'The tendency to get students to exchange
unmotivating, imposed information can even go to the extreme where
much of their 'communication' is about the behaviour of the fictional
characters of their coursebooks'.
Yet, like it or not, we live and work with grammatical syllabuses, and
with coursebooks whose characters may have only a slender relation to
the world which our learners inhabit. The challenge, then, from a
teacher training perspective, is this: how can teaching be made more
truly communicative, regardless of the constraints within which most
teachers work?

A 'bottom-up' Our approach, on Diploma courses, has been two-pronged. On the one
approach hand (and beyond the scope of this article; see, for example, Thornbury
1995), we explore approaches to lesson design that foreground the
learner's contribution to lesson content: this is the 'top-down' approach.
The 'bottom-up' approach, on the other hand, aims to raise teachers'
awareness of the discrete features of their classroom talk, through the
recording, transcription, and analysis of lesson sequences. The assump-
tion is that awareness is a prerequisite for change, and our philosophy is
consistent with Burns' (1990: 57) contention that 'if CLT is to become
more 'communicative'... teachers need to be encouraged to gain greater
understanding of the interactional processes of their own classrooms'.
I should add that no original claims are being made here as to the value,
from a training perspective, of the analysis of lesson transcripts. I am
aware that this is already an established practice and has been promoted
by Ramani (1987) and Nunan (1989), among others. My intention is
simply to demonstrate the effectiveness and practicability of this training
tool, for any trainer who might still be unconvinced, and to suggest how
it can be integrated into existing programmes.

Features of Our starting point is to establish a simple typology of communicative

communicative classroom discourse. (This can be done by having trainees participate in
classroom talk two mock 'mini-lessons', the one without, and the other with, the target
discourse features; alternatively, video extracts could be shown and
contrasted.)
Among the features that we feel are important to identify are the
following:

Referential Referential questions are genuine questions: those for which the teacher
questions does not know the answer, rather than display questions, whose primary
purpose is to allow the students to display their knowledge of language.
This is probably one of the features of teacher talk that has been most
thoroughly researched, and there is plenty of evidence (e.g. Long and
Sato 1983) to suggest that the vast majority of questions teachers ask are
display questions, whereas, in 'real life', of course, most questions are
referential. Nunan (1989: 30) suggests that 'it is not inconceivable that
Teachers research teacher talk 281
 the effort involved in answering referential questions prompts a greater
effort and depth of processing on the part of the learner'. I would go
further, and argue that the effort involved in asking referential questions
prompts a greater effort and depth of processing on the part of the
teacher. Try conducting a lesson in which every question is referential!
For teachers brought up in the 'elicit-standardize-drill' school, it can be
a salutary experience. Referential questions touch parts beyond the
reach of other types of question.

Content feedback Feedback on content involves responding to the content of what learners
are saying, rather than commenting solely on the form. After all, there is
not much point in asking referential questions if no attention is paid to
the meanings the learner is expressing. But ritualized responses, such as
'OK', irrespective of the message, anchor the classroom discourse firmly
in the traditional IRF camp, and suggest that 'it doesn't matter what you
say so long as you pronounce it properly'.

Wait time 'Wait time' is the time teachers allow students to answer questions,
before, for example, asking another student, rephrasing the question, or
even answering their own question themselves. Nunan (1991) quotes
studies showing that when teachers are trained to wait three or four
seconds, instead of the customary one, not only do more students
respond, but there is an increase in the average length of their responses.
The proportion of student-initiated questioning also increases. All of
these adjustments would seem to be worthy objectives in a commu-
nicative classroom.

Student-initiated If, as is argued by proponents of a communicative approach, acquisition

talk is facilitated by the negotiation of meaning in interaction, it follows that
learners should, at least some of the time, be asking the questions. A
high proportion of student-initiated questions would suggest a healthy
distribution of the 'ownership' of classroom discourse, which in turn
would tend to promote more 'investment' on the part of the learner: 'A
significant source of motivation and attention is lost when turn talking is
predetermined rather than interactionally managed by the participants!
(Van Lier 1988: 133). Slimani (1989) has also suggested that although
there is usually much less learner-initiated than teacher-initiated content
in classrooms, it is from the former, rather than the latter, that learners
claim to have learned the most.

Ritualized The above remarks do not imply that there is no place in the classroom
behaviours for display questions, rapid-fire IRF sequences, teacher-initiated talk,
and other examples of traditional classroom discourse. However,
teachers can get into bad habits, and one of the functions of in-service
training is to root out what Maingay (1988) calls 'ritualized behaviours'.
The extracts from trainees' analyses quoted below demonstrate that
awareness-raising, the first stage in de-ritualizing teaching, can be
achieved using simple classroom research procedures.
282 Scott Thornbury
Research project Having identified these features of communicative classroom talk,
trainees are asked to record, transcribe, and analyse a segment of
classroom talk (see Appendix for a suggested rubric). As preparation
they are recommended to read Nunan (1987), Burns (1990), and
Kumaravadivelu (1993). The finished analysis is submitted as an
evaluated project; more important, though, is the spin-off into actual
teaching practice, and the indications of increasing communicativeness
in the classroom which, while not always immediate, are invariably
rewarding.

Trainee comments As an indication of the kinds of insights the trainees reached, here is a
selection of their comments, grouped according to the four features of
communicative classroom talk described above:
Referential questions
My display questions prompted much shorter and less complex
answers than my referential questions . . . Perhaps it suggests that my
teaching style is too elicitation-focused.
From this extract I have also noticed that I only used nominated
questions on weaker students, who, I think, may have trouble with the
concepts, while the majority of open questions were answered by
stronger students. By limiting weaker students to concept [checking]
questions only, I may be reinforcing their insecurities by implying that
I think they don't understand.
Content feedback
At various points I am only reacting to how and not what is being
said.. .When I ask the question 'Do you listen to his music?' the
student responds 'I preferred him in the past. It like me more' and I
get her to correct herself. She then corrects herself voluntarily: 'I liked
it more, than the a c t u a l . . . er . . . present music'. I am so pleased with
this I practically shout 'Very good!'. However, I do not respond to
what she is saying, the obvious next question being 'Why did you like
his music better in the past?'. . .. -
'OK' was used, overused, and abused. I used it as feedback, to
punctuate my sentences, to signpost between stages. This is very
annoying, but I find my use of echoing much more worrying. I used
'echo' to model, remodel, I even echo myself, especially if students
don't come up with the answer immediately.
Wait time
One thing I think I have improved on and I felt listening to the
tape, is giving students sufficient waiting time and therefore them
having enough time to formulate an answer or a question. This I
have brought into my teaching since starting the Diploma and I
have noticed beneficial results for the students, including sponta-
neous interaction between students, and [between] me and the
students.
Teachers research teacher talk 283
 Student-initiated talk
I gave the students absolutely no space to do anything other than what
I wanted. Throughout the tapescript, it was the teacher versus 'them'
. . . I was the complete authority—ignoring some students' contribu-
tions in my rush to accomplish set aims. There was no interaction
between the students; I only allowed time for this in the free practice
at the end of the lesson, therefore the lesson was very divided. First
the teacher does all the talking, then you're on your own.

Metaphors Significant, perhaps, are the metaphors that the trainees use in
describing their somewhat mechanical, even bullying, style of teaching.
Some examples:
'I chose to bulldoze through most remarks . . . '
'The same objectives could have been achieved in a less robotic
fashion . . . '
'It seems that I was ploughing students through the furrows my lesson
was following . . . '
From a training point of view, the metaphors they use offer an insight
into trainee teachers' self-images, which in turn are an important locus
for change: without adjustments at the level of image, the effects of
training may only be superficial (see Thornbury 1991).

Transcripts Finally, here are transcripts of two lesson extracts, along with their
accompanying (edited) commentaries:
Extract 1
[An upper intermediate class has just read three texts about the musician
Phil Collins for comprehension purposes.]
T: OK, look at the last text on the sheet that Cathy gave you, OK?
. . . What's it about... the last text?
SI: The last text . . .
T: Who's it about?
S2: It's about Phil Collins' life.
T: Yeah. It's about Phil Collins . . . erm . . . what does Phil Collins
do?
S2: . . . singer.
SI: . . . plays drums I think.
T: He's a singer and he . . . ?
S3: Plays drums.
T: He's a singer and he plays the drums so he's a . . . ?
S4: Drummer, he's a drummer.
T: OK. Does he sing well? Does he sing well? Is he a good singer?
Ss: Yes. (laughter)
S5: No.
T: You think so? . . . Yeah, but you don't?
S5: No.
T: Is he a good drummer? . . . Does he play the drums well?
Ss: Yes. Yes.
284 Scott Thornbury
 T: Do you think that when he was a child he used to practise a lot?
Did he practise a lot?
Ss: Yes.
T: OK, where did he practise?
Ss: At home.
T: Did he practise a lot? . . . Did he practise a lot?
Ss: Yes . . . yes.
T: How do you know? . . . How do you know? . . . How do you
know?
SI: We read it before.
S4: Because he's very good.
T: So what does it say in the text?
S6: Always playing drums.
T: Yeah . . . again!
S6: He was always playing the drums.
T: He was always playing the drums, OK? He was always playing
the drums . . . Everybody . . . say it.
Ss: He was always playing the drums.
S2: Everyday.
S5: All the days . . . all the time.
T: Everyday .... . all the time yeah . . . always . . . good . . .

Teachers's comments
Virtually all the questions are display questions, which accounts for
the tiny- amount of 'real' communication and student talking-time.
Most of the questions are about something both the teacher and the
students have already read. The only examples that could possibly be
construed as being 'real' questions are 'Is he a good singer?' and 'Is he
a good drummer?'. Both require a personal opinion. However, the
text clearly considers Phil Collins to be a good singer and drummer
and so, in view of the type of questioning going on in the lesson, the
students' most natural reaction will be to come up with the opinion of
the text rather than to give their own . . . There are a couple of
moments when [real communication] threatens to break out, but these
are ruthlessly snuffed out by the teacher. The point when students
begin to differ in their opinions over the musical talents of Phil Collins
is not allowed to develop, but it could have lead to some more
authentic interaction between the students. [...] There is one occasion
when the teacher asks for information about the text and then
continues by asking 'How do you know?'. The teacher is expecting the
phrase from the text, but instead gets 'we read it before'. This could be
said to be real communication, because the student is perhaps tired of
the display questions, and so decides not to play the game any more.

Extract 2
[An elementary class has been working on the language of making
suggestions]
SI: What about go to mountains?
T: What about. . . ?
Teachers research teacher talk 285
 SI: What about going to mountains, we can do 'barrancking'.
[Ss laugh]
T: What's 'barrancking'?
S2: Is a sport.
T: Yes, but what do you do exactly?
S3: You have a river, a small river and . . . [gestures]
T: Goes down?
S3: Yes, as a cataract.
T: OK, a waterfall [writes it on board]. What's a waterfall, Manual?
Can you give me an example? A famous waterfall, [draws]
SI: Like Niagara?
T: OK. So what do you do with the waterfall?
S4: You go down.
T: What? In a boat?
S4: No, no, with a . . . IComo se dice cuerda?
S3: Cord.
T: No, rope, a cord is smaller, like at the window, look, [points]
S4: Rope, rope, you go down rope in waterfall.
S2: You wear? black clothes . . . /1 spejl klaudez /
T: / spejl ktauz / Repeat [student repeats] . . . [ . . . ] This sounds
dangerous, is it dangerous?
Ss: No, no.
S3: Is in summer, no much water.
T: Sorry?
S3: Poco . . . poco . . . little water, river is not strong.
T: OK . . . and you have done this? What's it called in Spanish?
S4: Barranquismo. In English?
T: I don't know. I'll have to ask somebody.
S2: It is good, you come? IComo es diu? Let's go together.
T: I don't think so. [laughs]
S4: Yes, yes, you come, we can go in summer.
T: Well, in the summer, not now, it's too cold.
Ss: No, no.
Teacher's comments
I chose this particular extract to transcribe because I feel it is an
example of genuine interaction between myself and the students, and
in particular because! consider it a breakthrough point with this class.
Being an elementary class . . . the students, up till this point, had
displayed reluctance^m-initiatingidiscussions,_ askingijne-questionsr
etc., probably due to insecurity about their"language-ability and,.in-
terms of classroom dynamics, the lack of any one particular student
who is prepared to act as a catalyst and 'break the ice' . . .
During the class . . . a genuine 'information gap' occurred, as the
students tried to explain a sport I am completely unfamiliar with.
Consequently, insecurities were forgotten as the students enthusias-
tically attempted to communicate the concept, and they surprised
themselves (when I replayed the tape) with the volume of language
produced . . .
286 Scott Thornbury
 In general I tend not to interact enough with students, having believed
that the students should be encouraged to work with each other as
much as possible and the teacher should take a back seat role. I still
believe that this is a good idea, in the right circumstances—group/pair
work, etc.—but I think I have underestimated the value of interaction
with the only native speaker the students have access to—their
teacher.
I hope it is apparent from Extract 2 that these assignments are not
always exercises in self-mortification. And that, by raising awareness as
to what constitutes truly communicative classroom discourse, break-
throughs in this area are possible. It should be emphasized, however,
that, as in all things, change in teaching behaviours is neither painless
nor linear. Reducing an over-reliance on IRF sequences can leave many
teachers feeling disempowered and 'un-teacherlike'. The washback
effect of institutionalized teaching assessment schemes means that
change is sometimes short-lived. Nor do adjustments to teacher talk
have much long-term effect if the teacher is not committed to the belief
that student-centredness is more than a matter of providing pair and
group work, and getting the students to ask the questions. The training
programme must also address the larger concern of the learner's
personal investment in the language learning process. Although that
concern is beyond the scope of this article, it can be said that the critical
examination of classroom discourse offers a convenient tool for the
further exploration of these issues.
Received December 1995

Teachers research teacher talk 287

Acknowledgement ing: making it work'. ELT Journal 41/2:136-45.
I would like to thank participants on the RSA/ Nunan, D. 1989. Understanding Language Class-
Cambridge Diploma TEFLA courses at Interna- rooms. Hemel Hempstead: Prentice Hall.
tional House, Barcelona, from whose projects I Nunan, D. 1991. Language Teaching Methodol-
have quoted, and my colleague Neil Forrest, ogy. Hemel Hempstead: Prentice Hall.
whose help in elaborating this 'thesis' has been Palmer, H. E. 1940, 1970. The Teaching of Oral
invaluable. English. London: Longman.
Ramani, E. 1987. 'Theorizing from the classroom'.
References ELT Journal 41/1: 3-11.
Bums, A. 1990. 'Focus on language in the Slimani, A. 1989. 'The role of topicalization in
communicative classroom' in G. Brindley (ed.) classroom language learning'. System 17:
The Second Language Curriculum in Action. 223-34.
Sydney: NCELTR. Swan, M. 1985. 'A critical look at the commu-
Kumaravadivelu, B. 1993. 'Maximizing learning nicative approach (1) and (2). ELT Journal
potential in the communicative classroom'. ELT 39/1: 2-12; ELT Journal 39/2: 76-87.
Journal 47/1: 12-1. Republished in R. Rossner and R. Bolitho
Legutke, M. and H. Thomas 1991. Process and (eds.) 1990 Currents of Change in English
Experience in the Language Classroom. Harlow: Language Teaching. Oxford: Oxford University
Longman. Press.
Long, M. and C. Sato 1983. 'Classroom foreigner Thornbury, S. C. 1991. 'Metaphors we work by:
talk discourse: forms and functions of teachers' EFL and its metaphors'. ELT Journal 45/3:
questions' in H. Seliger and M. Long (eds.) 193-200.
Classroom Oriented Research in Second Thornbury, S. 1995. 'Not waving but drowning:
Language Acquisition. Rowley, Mass.: Newbury perils of the grammar revival'. Paper presented
House. at a teacher training conference, International
Maingay, P. 1988. 'Observation for training, House, London.
development, or assessment?' in T. Duff (ed.) Van Lier, L. 1988. The Classroom and the
Explorations in Teacher Training: Problems and Language Learner. Harlow: Longman.
Issues. Harlow: Longman. Willis, D. 1990. The Lexical Syllabus. London:
Nunan, D. 1987. 'Communicative language teach- Collins ELT.

288 Scott Thornbury

 Appendix Project: Teacher Talk
Task rubric
Aim To investigate the way you conduct classroom interactions.
Method Record and transcribe a teacher-fronted, lock-step segment
of one of your lessons. It should not be too long—about a
page of transcription should be enough. If you like, you
could analyse several shorter segments. Use conventions
such as:

T = teacher
SI = a student
S2 = another student
Ss = students altogether
(inaud) = inaudible
Analysis 1. Identify the different teacher 'moves' in the transcript—
for example:
explaining
eliciting
giving feedback (including correcting)
modelling
checking understanding
conversing
2. Categorize the questions in the extract. For example,
are they:
nominated or open?
teacher-initiated or student-initiated?
display questions or 'real' questions?
grammatical or ungrammatical?
3. Identify any examples of 'real' communication between
teachers and students or students and teachers.
4. Identify any examples of communication breakdown.
How do you account for them?
5. Identify any vocal 'mannerisms' you may have, e.g.
ritualistic responses to student utterances, such as
'Good' or 'OK?'.
6. Finally, evaluate the interaction overall. What does it
suggest about your teaching style? What changes might
you like to implement in your classroom talk?

The author Reading. He has written ELT materials for

Scott Thornbury directs the teacher training Longman and Heinemann, and is the author of a
department at International House, Barcelona forthcoming book of language analysis tasks,
and has an MA TEFL from the University of About Language (Cambridge University Press).

Teachers research teacher talk 289