‘Small Talk’: developing fluency,

accuracy, and complexity in speaking

James Hunter

Downloaded from https://academic.oup.com/eltj/article/66/1/30/391278 by Glasgow University Library user on 15 May 2023

A major issue that continues to challenge language teachers is how to ensure that
learners develop accuracy and complexity in their speaking, as well as fluency.
Teachers know that too much corrective feedback (CF) can make learners
reluctant to speak, while not enough may allow their errors to become entrenched.
Furthermore, there is controversy over the effectiveness of recasts (the most
common form of CF) in promoting acquisition. This article explores
a methodology, ‘Small Talk’, which aims to resolve some of the tensions between
the need to encourage truly communicative language use and the need to develop
complexity and to bring focus on forms into the syllabus in ways that can be
recognized as valid and relevant by both teachers and learners. It presents some
preliminary research on the viability of this CFmethodology premised on attention
to, and arising from the needs of, the individual learner.

Introduction A perennial struggle for teachers is how to develop both accuracy and
fluency in students’ speaking since one often seems to come at the expense
of the other. On top of this, we have the even greater challenge of coaxing our
students out of their comfort zones towards greater complexity (Skehan
1998), especially when the language they have appears to be adequate for
their communicative purposes. Different theoretical positions have had
dramatic and conflicting influences on teaching methodology, so it is not
always clear what we should be doing to best serve our students. If they
practise pronunciation, vocabulary, and grammar, will they use them
spontaneously and correctly when necessary? Should we teach grammar
explicitly, and if so, which forms should we teach? Should we correct errors,
and if so, how, and which ones? The wise teacher employs an eclectic
combination of methods depending on the teaching context and the
students in the classroom, but it is hard to escape the feeling that eclectic
often simply means unsystematic.

The limitations of Many teachers resist the strong form of communicative language teaching
contemporary (CLT) because it does not have ‘concrete’, ‘tangible’ content and, therefore,
language teaching does not equate with ‘real’ teaching. This is hardly surprising since the one
area in which language teachers have traditionally had expertise, the
structure of the language, is off-limits in the strong form of CLT; all that
remains is coaching learners on how to get their message across, which in
the final analysis can be done with very limited linguistic resources,

30 E LT Journal Volume 66/1 January 2012; doi:10.1093/elt/ccq093

ª The Author 2011. Published by Oxford University Press; all rights reserved.
Advance Access publication March 15, 2011
 provided that formal accuracy is not a major concern or a concern at all.
Indeed, Gatbonton and Segalowitz (2005: 327) see ‘no provisions in current
CLT methodologies to promote language use to a high level of mastery
through repetitive practice’, noting that ‘focused practice continues to be
seen as inimical to the inherently open and unpredictable nature of
communicative activities’. Thus, while we can fairly assume that a teacher-
centred classroom in which the main focus is on linguistic form will not lead
to fluency, we can also be confident that a focus on authentic
communication alone will not lead to accuracy and complexity.

Downloaded from https://academic.oup.com/eltj/article/66/1/30/391278 by Glasgow University Library user on 15 May 2023

It could be that the pendulum will return towards pedagogy that prioritizes
formal accuracy over communicative fluency, but I doubt this for several
reasons. First, sociolinguistic research into language varieties has
challenged the notion that there is a monolithic, ‘correct’ form—that of the
‘native speaker’—against which the language of learners can be measured.
Second, this challenge has increased pressure on researchers, materials
writers, and teachers to check their linguistic intuitions against findings
from corpus linguistics, which continue to shed light on the importance of
context at both the linguistic and sociolinguistic level. Finally, language-
teaching methodologies have become increasingly humanistic, stressing
the importance of the learner in the language acquisition process. The
heterogeneity of linguistic competence, learning styles, strategies, and
degree of social investment of language learners is precisely the impetus for
greater research efforts into pedagogical methodologies that depart from
the prescriptive syllabus and encourage our reflective and intuitive capacity
as teachers. The time is right for a responsive pedagogy premised on careful
attention to, and arising from the needs of, the individual learner.

The origins of Brumfit (1979) was the first to highlight the distinction between fluency,
accuracy and fluency which represents the learner’s ‘truly internalized grammar’, contrasting
this with ‘overt and conscious accuracy’ (115, emphasis in original) and
suggested that fluency should be ‘regarded as natural language use,
whether or not it results in native-speaker-like language comprehension or
production’ (Brumfit 1984: 56). When he introduced these terms as key
concepts in second language acquisition (SLA) and syllabus design, Brumfit
was also arguing for an approach to form- and meaning-focused teaching,
which, it seems, has largely fallen on deaf ears. For instance, he proposed
allowing people to operate as effectively as they [can], and attempting to
mould what they [produce] in the desired direction, rather than explicitly
teaching and expecting convergent imitation. (ibid.: 50)
That is, instead of giving learners language items to imitate and expecting
their imitations gradually to conform to the model, teachers could discover
what learners actually wanted to say and then teach them how to say it in the
target language. None the less, it is still rare to leave learners to their own
devices to produce ‘natural language use’, partly owing to the fear of
exposing students to each others’ errors, but also because in many
classrooms students rarely have extended opportunities to produce
language for themselves at all. Rarer still is the learner-driven syllabus that
Brumfit proposed, one in which teaching is based on language production,
and not the other way around.

Developing fluency, accuracy, and complexity in speaking 31

The role of corrective Corrective feedback (CF) literature to date has, with very few exceptions,
feedback in oral examined feedback provided by teachers during teacher-fronted activities, in
fluency activities which the teacher controls the activity itself as well as the type and quantity
of language produced. Research into the effects of such feedback has
tentatively suggested a positive role for CF in the form of recasts, but has
been weakened by methodological issues such as the interpretation of
teacher intent and learner perception of recast moves (Mackey, Al-Khalil,
Atanassova, Hama, Logan-Terry, and Nakatsukasa 2007) and the
controversy over learner uptake as an indicator of either noticing or actual

Downloaded from https://academic.oup.com/eltj/article/66/1/30/391278 by Glasgow University Library user on 15 May 2023

acquisition. Whether or not recasts are the most effective form of CF (see
Ammar and Spada (2006) for a contrasting view) the pedagogical goal
remains, to return to Brumfit, ‘convergent imitation’.
What would an alternative pedagogy and CF approach look like? Brumfit
(1979: 115), talks of the teacher modifying the learners’ ‘self-developed
systems as reflected in the fluent language behaviour’ claiming that
teachers ‘need to look at genuine language use in the classroom, to the
extent that it can ever be really genuine’ (Brumfit 1984: 52). But this
presupposes two conditions: fluent—and genuine—language behaviour
and a way to encourage learners to focus on the formal aspects of their
production. Skehan (1996) suggests that these are unlikely to occur
simultaneously since students engaged in genuine communicative
interaction are likely to be too focused on meaning to pay attention to form.
The same must be said of teachers; however, it is extremely difficult to
participate in, let alone direct, a genuinely communicative interaction while
simultaneously paying attention to and remembering the form of the
utterances produced. Therefore a third condition is that teachers be free to
listen carefully to both form and content of student utterances, which
means being free from the responsibility to direct or even to participate in
the interaction. This would permit teachers to become the experts on the
language their students actually use and to design effective pedagogies to
help them progress; and it would bring much-needed content to C LT and
highly relevant content at that. The ‘real teaching’ that teachers feel is
currently missing would be the language that the learners are striving for at
that moment, rather than the syllabus imposed by textbooks, which is
disconnected from the needs of the learner at best, and completely arbitrary
at worst. And finally, since language learning occurs over time and learners’
‘self-developed systems’ are likely to change at different rates, it is essential
that the CF methodology be responsive to the needs of the individual learner
and that there be some systematic means of collecting, storing, analysing,
and recalling the data collected.

The communicative ‘Small Talk’ began as an experiment in learner-centred, reflective teaching

methodology: of oral communication over 20 years ago (Harris 1998) and has developed
‘Small Talk’ into a comprehensive approach to developing accuracy, fluency, and
complexity in oral production. In a ‘Small Talk’ session, students use their
communicative ability in conversation without intervention by the teacher,
and then receive feedback. Each session has a pre-appointed student leader,
who is responsible for choosing the topic, providing questions and relevant
vocabulary to further the discussion, putting classmates into small groups,
timing the conversation, and leading a ‘check-in’ session at the end, in

32 James Hunter
which each group reports to the whole class on their conversation. The
stages and timing of a ‘Small Talk’ session are usually similar to that shown
below.
1 The day before the session, the leader announces the
topic.
2 At the beginning of the session, the leader writes (3–5 minutes)
discussion questions and vocabulary on the board,
re-introduces the topic, and clarifies any confusion;
the leader also puts the students into groups of three

Downloaded from https://academic.oup.com/eltj/article/66/1/30/391278 by Glasgow University Library user on 15 May 2023

to four and tells the students to begin.
3 Groups discuss the topic. (15–20 minutes)
4 The leader asks the groups to bring their (5 minutes)
conversation to a close and prepare for check in; the
groups decide what to report to the class and who will
do it.
5 The leader invites each group to check in with the (5–10 minutes)
class about the highlights of their conversation.
6 The leader thanks the class and reminds them of the (1 minute)
next ‘Small Talk’ date and leader.

The students are encouraged, in Stages 4 and 5 above, to reflect and report
on the dynamics of their interaction and their own part in it. This makes
explicit the quality of conversational interaction as both a cultural construct
(i.e. different cultures ‘do’ conversation in very different ways) and
a quantifiable variable (i.e. we can identify the features of appropriate
interaction and evaluate our use of them).
The teacher, having no role in or responsibility for the conversations, is able
to observe the interactions and afterwards to suggest ways in which they can
be improved. In a typical 50-minute class, there are usually ten minutes at
the end for ‘coaching’, when the teacher comments on the interaction and
dynamics of the ‘Small Talk’ session. For instance, I often teach or remind
quiet or non-fluent students ways to get their point across; I remind
dominating talkers to be patient and to invite others to participate; and we
practise how to ‘listen actively’, to show interlocutors our comprehension (or
lack of it) and to interrupt for clarification whenever necessary. ‘Small Talk’
is thus effective in increasing the students’ pragmatic competence since it
gives them an opportunity to practise, in a relatively low-stress environment,
the kinds of speech acts they would need in higher stress interactions
outside the classroom. It also puts students in the position, as leaders, to
practise a variety of speech acts and discourse management strategies that
are usually restricted to the teacher.
‘Small Talk’ is very popular with students, as the following comments (from
end-of-semester class evaluations) illustrate:
n its helped me in my speaking a lot.
n i think it will improve our skills.
n i really enjoy it because we chose our topic.
n i recommend it for student.
In addition, at least from teachers’ untested observations, it is very effective
in raising the level of fluency of lower-intermediate to advanced students in

Developing fluency, accuracy, and complexity in speaking 33

 general and particularly of students from cultural backgrounds in which
verbal participation is not encouraged. It is not suitable for true beginners,
who do not have sufficient language for what Willis (2003: 22) calls
‘improvisation’, in which ‘learners are obliged to make the most of the
language they have at their command’. For them, perhaps more appropriate
would be what Willis calls ‘consolidation’ activities, in which ‘learners think
through carefully what they want to say’, which would more accurately
describe most classroom tasks. However, even in improvisation activities
(and perhaps especially then), students understandably want to know what

Downloaded from https://academic.oup.com/eltj/article/66/1/30/391278 by Glasgow University Library user on 15 May 2023

they are not doing successfully, and ‘Small Talk’ also gives an opportunity for
teachers, as observers of their students, to focus on accuracy.

‘Small Talk’ Since the goal is for teachers not to intrude in the conversation with
worksheets comments, recasts, or other corrective moves, CF is provided in the
following way. It would be impossible to listen to four or five conversations
(or however many groups there are) simultaneously, but teachers can catch
a portion of each conversation, listening to each group in turn and writing
down inaccurate language use, whether it interferes with the
communicative flow or not. They then enter each error (typically 15 to 50 per
‘Small Talk’ session) with the name of the speaker into a computerized
database,1 noting the date of the ‘Small Talk’ session and the topic (Figure 1).

fi g u r e 1
Worksheet entry form
from the database

34 James Hunter
 Downloaded from https://academic.oup.com/eltj/article/66/1/30/391278 by Glasgow University Library user on 15 May 2023
fi g u r e 2
Excerpt from a ‘Small
Talk’ worksheet

Teachers also occasionally flag an item for all students to correct, regardless
of who said it, which allows them to focus on specific language points. This
option is especially useful in cases where several students are making
similar kinds of errors. The database produces a worksheet of these errors
(Figure 2), which is normally made available to the students within 24 hours
of the conversation.
If certain individuals dominate the conversations, of course, this collection
of errors would be biased towards those individuals and some students
would rarely be heard by the teacher. Consequently, two mechanisms are in
place to counteract this effect. First, as mentioned above, the teacher
addresses domination during the coaching sessions and explicitly teaches
discourse strategies to reduce it. Second, because the database keeps
a running tally of the speakers and their errors, it is possible to form groups
consisting of individuals who have not been heard as frequently (and who
often tend to be quieter and less dominant) and spend more time (even the
whole session, if necessary) listening exclusively to them.

CF options Giving learners a written transcription of their errors enables them to

correct any ‘slips’ they have made, and it might push them towards a more
stable interlanguage form in cases where there is variability, and this alone
makes the activity worthwhile. However, beyond that, if learners truly do not
know how to say something because they lack the appropriate structure or
vocabulary, some form of guidance is necessary to facilitate more accurate
production in the future. Two choices present themselves. The first is to
provide the students with some sort of written metalinguistic feedback to
enable them to locate and correct the error (Figure 3). This option has
intuitive appeal and widespread support in the literature, especially the
literature on feedback in writing (Ellis 2009).
The second option is to provide the students with the printed worksheet of
errors along with reformulated versions, as a competent speaker might say
them, in the form of an audio recording. Students then listen to this in order
to work out where the differences lie. As in a dictation, students have to
listen very carefully to hear some of the less salient grammatical features (in
particular, articles and verb inflections). Furthermore, this option gives
teachers the opportunity to introduce alternative, often more complex,
language forms that can express the students’ intended meanings and has

Developing fluency, accuracy, and complexity in speaking 35

 Downloaded from https://academic.oup.com/eltj/article/66/1/30/391278 by Glasgow University Library user on 15 May 2023
fi g u r e 3
Worksheet with
metalinguistic feedback

the additional advantage of providing a correct model for phonological

errors.
Pedagogically, both these options satisfy teachers’ concerns that students
actually do something with the CF, and in theoretical terms, the hypothesis
is that this level of focus is more likely to lead to acquisition than the ‘uptake’
of simply repeating a teacher’s recast. Since this is delayed CF, there is no
immediate communicative need for the information, the moment has
passed. However, it might better help students to ‘notice the gap’ (Schmidt
and Frota 1986) because there is no simultaneous pressure to
communicate. It also constitutes both explicit positive evidence and implicit
negative evidence about the language (Long, Inagaki, and Ortega 1998).
However the feedback is provided, the students keep a running list of their
own errors and errors which the teacher has flagged, on which they are
tested every three to four weeks. The test requires them to look at these
errors and orally correct as many as they can in a given amount of time,
usually two or three minutes. As an example, the following sentences were
taken from a conversation about ‘Traditional and modern culture’ from my
class of 22 adult intermediate students (L1 Arabic):
* We can learn what their food, their cultures.
* In the past the womans wear the traditional clothes.
* Yeah, actually I’m agree with you.
When I tested the students on these sentences (and many others) six weeks
later, all 22 could fluently produce correct forms, typically:
n We can learn what their food and their cultures are.
n In the past women wore traditional clothes.
n Yeah, actually I agree with you.
I do not claim that all the students had acquired all or any of the previously
incorrect forms, and therefore that they would be able to produce the correct
form fluently in novel contexts; but the focus on these forms did have the
noticeable effect of promoting self-correction, especially of high-frequency
chunks such as *‘ . . . I’m agree . . .’, in subsequent ‘Small Talk’ sessions
without any reduction in their overall willingness to speak. On the contrary,

36 James Hunter
 the students welcomed the individualized attention to their spoken
production and felt more willing to try to express themselves knowing that
I would be listening and providing feedback, as the following comment
shows:
I want to liste all of my sentences during small talk. I am not used to speak
correct sentenses. So I am often surprised at seeing my mistakes. If I can
get more sentenses, I can edit my sentenses more.

Downloaded from https://academic.oup.com/eltj/article/66/1/30/391278 by Glasgow University Library user on 15 May 2023

Questions I made a small-scale study of the CFpotential of ‘Small Talk’ with a class of 12
adult intermediate students (mixed L1) in an academic ESL program in the
United States. Ten of the weekly ‘Small Talk’ sessions were videotaped (see
the Appendix), and four of these were randomly selected for analysis. The
conversations were transcribed and turns with errors were identified. I then
asked five experienced teachers to watch the videos independently, without
stopping or rewinding, and make worksheets just as I (the class teacher) had
done during the sessions. In doing this, I wanted to address the following
questions:
1 Do students get more speaking practice during ‘Small Talk’ than during
a traditional, teacher-fronted class? Do they make more errors?
2 What percentage of students’ errors receives CF, and what percentage of
uptake is there?
3 Do some students receive more CF than others, and if so, why?

Results In answer to the first question, the results from the four ‘Small Talk’
transcripts are shown in Table 1.

Topic Time Word count Turns Errors % of turns with errors

Favourite place 31:53 1,756 308 87 28
Traditional food 33:32 2,795 326 111 34
table 1
$1 Million 32:35 2,723 344 95 28
Count of words, turns,
Generation gap 26:20 2,696 279 106 38
and turns with errors in
Total 124 9,970 1,270 399 31
four transcripts

In this study, there were 1,270 student turns in 124 minutes of conversation;
by way of comparison, the oft-cited study by Lyster and Ranta (1997: 52 and
62) documented 3,268 student turns in 1,100 minutes. Lyster and Ranta do
not include word counts, but in turn count alone the students in ‘Small Talk’
spoke 3.5 times more than those in Lyster and Ranta’s study. The percentage
of student turns with errors in both studies is almost the same, 31 per cent in
this study and 34 per cent in Lyster and Ranta’s (ibid.: 52), meaning that the
speakers left to their own devices not only spoke more but also made slightly
fewer errors than those in teacher-controlled activities.
To address the second question, the number of erroneous utterances written
down by each teacher over four ‘Small Talk’ sessions was calculated as
a percentage of the number of student errors identified in the transcripts
(Table 2).

Developing fluency, accuracy, and complexity in speaking 37

 Teacher Mean

table 2 T1 T2 T3 T4 T5 T6
Percentage of errors
34% 46% 36% 57% 42% 24% 40%
identified by teachers
T1 was the class teacher.

The level of error identification by the teachers ranged from 24 per cent to 57

Downloaded from https://academic.oup.com/eltj/article/66/1/30/391278 by Glasgow University Library user on 15 May 2023

per cent, giving an overall average of 40 per cent. The figure of 34 per cent
for T1 is the percentage of all errors from these four sessions that I actually
provided to these students as CF. Even the low figure here (T6’s 24 per cent)
would probably be acceptable: if students knew that even a quarter of their
errors would be identified by their teacher, they would certainly not feel that
they were wasting time, let alone if they could be confident that around
40 per cent of the errors were being targeted. A comparison can again be
made with the study by Lyster and Ranta (1997: 53), in which ‘of the total
number of errors produced by students, only 17 per cent of errors eventually
lead to repair’; in ‘Small Talk’, an average of 40 per cent of student errors
would eventually lead to uptake and repair since the students have to correct
the worksheets.
In addressing the third question, the number of errors produced by each
student (as identified in the transcripts) was compared to the number for
each student on the teacher worksheets. In addition, I calculated the
number of errors for each student that I (T1) identified over the entire
semester, in other words the amount of CF that the students actually
received over 16 weeks, giving a point of comparison for bias (Figure 4).
(Three students who were not present for the entire semester, S2, S4, and
S11, have been excluded from this analysis.)
The correlation between number of errors for each student found in the
transcripts and numbers of errors for each student appearing on the
worksheets of teachers was high, at .89. It is possible, of course, given the
random sampling procedure (the students were grouped by the leader, the
video recorded only a five- to seven-minute portion of each group’s
conversation), that some students would feature more than others and
therefore that the teacher identification of errors would be skewed more
towards them than others. It turns out, however, that all teachers identified
more errors for students who were more inaccurate overall, regardless of
how much they spoke. In other words, the CF provided closely reflected the
needs of individual students.

Conclusion We frequently tell our students that it is okay to make mistakes and that they
will not make progress unless they talk more. However, we also frequently
complain about the number of ‘basic’ errors that our students make. Willis
(2003) reminds us that this is both inevitable and desirable: errors are part of
the developmental process, and ‘it is the learners’ attempts to mean that pave
the way for learning’ (ibid.:110–111, emphasis added) and for noticing what
they need to learn. While some might argue that allowing students at an
intermediate or lower level to ‘improvise’ in the classroom could lead to
linguistic anarchy, I agree with Willis that opportunities for improvisation in

38 James Hunter
 Downloaded from https://academic.oup.com/eltj/article/66/1/30/391278 by Glasgow University Library user on 15 May 2023
fi g u r e 4
Identification of errors for
each student in
transcripts and by
teachers

the classroom are essential. Although space does not permit an analysis of
the discourse structure of the conversations, the transcripts show, as Willis
(1992) notes, that ‘in the absence of the teacher, [students’] interaction
becomes far richer’ (ibid.:180).
However, without some consistent way of observing and recording these
‘attempts to mean’, interpreting them, teaching to them, and assessing
subsequent learning, the teaching syllabus remains largely arbitrary and
disconnected from the needs of the learner. ‘Small Talk’ is a consistent
methodology for analysing and responding to learner language, and it
appears to target learners differentially in response to their self-developed
systems. It compares very favourably with the study of Lyster and Ranta
(1997) of CF in terms of the quantity of student interaction and CF provided.
Finally, my research indicates a connection between this methodology and
the development of accuracy, complexity, and fluency, and I am currently
looking at ways to evaluate the nature and strength of this connection.

Developing fluency, accuracy, and complexity in speaking 39

 Final revised version received December 2010

Note Mackey, A., M. Al-Khalil, G. Atanassova, M. Hama,

1 A self-contained version of the database is A. Logan-Terry, and K. Nakatsukasa. 2007. ‘Teachers’
available for download at http:// intentions and learners’ perceptions about corrective
www.gonzaga.edu/tesolresearch feedback in the L2 classroom’. Innovation in
Language Learning and Teaching 1/1: 129–52.
Schmidt, R. and S. Frota. 1986. ‘Developing basic
References conversational ability in a second language. A case

Downloaded from https://academic.oup.com/eltj/article/66/1/30/391278 by Glasgow University Library user on 15 May 2023

Ammar, A. and N. Spada. 2006. ‘One size fits all? study of an adult learner of Portuguese’ in R. Day
Recasts, prompts, and L2 learning’. Studies in Second (ed.). Talking to Learn: Conversation in Second
Language Acquisition 28/4: 543–74. Language Acquisition. Rowley, MA: Newbury House.
Brumfit, C. J. 1979. ‘Notional syllabuses—a Skehan, P. 1996. ‘A framework for the
reassessment’. System 7/2: 111–6. implementation of task-based instruction’. Applied
Brumfit, C. J. 1984. Communicative Methodology in Linguistics 17/1: 23–62.
Language Teaching: The Roles of Accuracy and Fluency. Skehan, P. 1998. A Cognitive Approach to Learning
Cambridge, UK: Cambridge University Press. Language. Oxford, UK: Oxford University Press.
Ellis, R. 2009. ‘A typology of written corrective Willis, D. 2003. Rules, Patterns, and Words: Grammar
feedback types’. ELT Journal 63/2: 97–107. and Lexis in English Language Teaching. Cambridge,
Gatbonton, E. and N. Segalowitz. 2005. ‘Rethinking UK: Cambridge University Press.
communicative language teaching: a focus on access Willis, J. 1992. ‘Inner and outer: spoken discourse in
to fluency’. The Canadian Modern Language Review the language classroom’ in M. Coulthard (ed.).
61/3: 325–53. Advances in Spoken Discourse Analysis. London:
Harris, R. 1998. ‘Making grammar instruction Routledge.
relevant through student-run conversations’. Paper
presented at the T E S O L convention, Seattle, The author
Washington. James Hunter has been teaching in the E S L and
Long, M. H., S. Inagaki, and L. Ortega. 1998. ‘The M AT E SL programs at Gonzaga University, in
role of implicit negative feedback in SLA: models Washington State, for the past ten years and is
and recasts in Japanese and Spanish’. The Modern currently completing his PhD in Applied Linguistics
Language Journal 82/3: 357–71. at the University of Birmingham, investigating
Lyster, R. and L. Ranta. 1997. ‘Corrective feedback corrective feedback in language teaching and C AL L.
and learner uptake: negotiation of form in He has taught in Spain, Japan, and most recently in
communicative classrooms’. Studies in Second Abu Dhabi.
Language Acquisition 19: 37–66. Email: hunter@gonzaga.edu

40 James Hunter
Appendix
Small Talk sessions, Level Date Worksheet no. Topic
Level 105/6, Spring
2008 105 23 January 2008 1 Sports
105 28 January 2008 2 Childhood
105 4 February 2008 3 Favourite place (31:53)
105 12 February 2008 4 Celebrations

Downloaded from https://academic.oup.com/eltj/article/66/1/30/391278 by Glasgow University Library user on 15 May 2023

105 18 February 2008 5 Dancing and parties
105 21 February 2008 6 Traditional food (33:32)
105 26 February 2008 7 How to look after your
body
106 25 March 2008 1 Cohabitation
106 27 March 2008 2 Crime and punishment
106 1 April 2008 3 Your dream
106 8 April 2008 4 If you had $1 million
(32:35)
106 15 April 2008 5 Conflict
106 17 April 2008 6 Discussion of novel
Whirligig
106 22 April 2008 7 Teamwork
106 24 April 2008 8 Generation gap (26:20)
106 29 April 2008 9 One day left on Earth
Bold items represent videotaped sessions

Developing fluency, accuracy, and complexity in speaking 41