International Journal of Bilingual Education and

Bilingualism

ISSN: 1367-0050 (Print) 1747-7522 (Online) Journal homepage: https://www.tandfonline.com/loi/rbeb20

How much L1 is too much? Teachers' language use

in response to students’ abilities and classroom
interaction in Content and Language Integrated
Learning

Yuen Yi Lo

To cite this article: Yuen Yi Lo (2015) How much L1 is too much? Teachers' language use in
response to students’ abilities and classroom interaction in Content and Language Integrated
Learning, International Journal of Bilingual Education and Bilingualism, 18:3, 270-288, DOI:
10.1080/13670050.2014.988112

To link to this article: https://doi.org/10.1080/13670050.2014.988112

Published online: 13 Dec 2014.

Submit your article to this journal

Article views: 4120

View related articles

View Crossmark data

Citing articles: 23 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=rbeb20
International Journal of Bilingual Education and Bilingualism, 2015
Vol. 18, No. 3, 270–288, http://dx.doi.org/10.1080/13670050.2014.988112

How much L1 is too much? Teachers’ language use in response to

students’ abilities and classroom interaction in Content and Language
Integrated Learning
Yuen Yi Lo*

Faculty of Education, University of Hong Kong, Hong Kong, China

(Received 30 April 2014; accepted 5 September 2014)

In Content and Language Integrated Learning (CLIL) classrooms where students’ L2

proficiency has not reached the threshold level, teachers have been observed to use L1
to assist students in grasping specific technical terms and abstract concepts. It is
argued to be a ‘realistic’ approach to the learning problems caused by students’ limited
L2 proficiency, particularly in English-as-a-foreign-language contexts. Nonetheless,
researchers have also warranted ‘judicious’ and ‘principled’ use of L1 so that both
content and language learning are facilitated. The question thus remains is whether
teachers can use L1 appropriately to suit their students’ needs. This study seeks to
address this question. It includes data from 30 Grade 10 lessons in Hong Kong CLIL
classrooms, where students’ L2 proficiency varied considerably. Using both quantit-
ative and qualitative analyses, this study shows that when teaching students with
limited L2 proficiency, teachers used a significant proportion of L1 in lessons to
explain the subject content, interact with students and develop students’ L2
metalinguistic awareness. In contrast, with students highly proficient in L2, teachers
used little L1, mainly to provide translation equivalents for L2 subject-specific
vocabulary items. This study thus shows that teachers seemed to be sensitive when
making use of students’ existing linguistic repertoires.
Keywords: content and language integrated learning; classroom interaction; code-
switching; bilingual education

Introduction
Content and Language Integrated Learning (CLIL1) is defined as ‘an education approach
in which various language-supportive methodologies are used which lead to dual-focused
form of instruction where attention is given both to the language and the content’ (Coyle,
Hood, and Marsh 2010, 3). Such a definition of CLIL highlights one very important
feature – the dual goal of both content and L2 learning. The stated aim of CLIL is that
students will benefit in their L2 development while being on par with those studying in
non-CLIL programmes in their academic achievement.
Although CLIL can theoretically target at any language, in recent decades, English as
an international language is usually chosen in CLIL in many English-as-a-foreign-
language (EFL) contexts such as the European continent and Asian countries (where
CLIL is implemented in the form of English-medium (EMI) education; Dalton-Puffer,
Nikula, and Smit 2010). In these contexts, the potential socio-economic benefits

*Email: [email protected]

© 2014 Taylor & Francis

 International Journal of Bilingual Education and Bilingualism 271

associated with English command result in an overwhelming preference for CLIL among
parents and students. Nonetheless, many students do not reach the threshold level of
English proficiency to benefit from the programmes (Lin 2012). Consequently, while they
enjoy certain benefits in L2 learning, they may lag behind in academic and cognitive
development, when compared with their peers learning through the first language (L1; Lo
and Lo 2014). To assist students in overcoming the language barrier when learning
academic content, teachers have been observed to use students’ first language (L1) in
CLIL lessons (Wannagat 2007).
The use of L1, or code-switching,2 has long been a controversial issue in the field of
second language acquisition. As one major goal of CLIL concerns L2 learning, the use of
L1 in CLIL lessons is also debatable. In early years, some researchers maintained the
immersion principle and oppose using L1 in bilingual programmes (Swain 1986). In
recent years, there seems to be a certain degree of consensus among researchers on the
functions of L1 in facilitating classroom interaction, and helping students to learn abstract
concepts and even L2 (Turnbull and Dailey-O’Cain 2009). Despite such recognition of
the values of L1, there are still concerns about its practice, i.e. teachers may use ‘too
much’ L1 injudiciously, which may in effect undermine the effectiveness of CLIL. Using
Hong Kong as a context, this paper thus seeks to examine the extent to which teachers
use L1 to achieve the dual goals of CLIL.

Literature review
The prevalence of CLIL and its challenges
Starting from the Canadian immersion programme, using the target language as the
medium of instruction (MOI) has been an appealing alternative to traditional isolated
language classes to facilitate L2 learning. While the immersion programme has been
found to put less emphasis on form-focused instruction (Lyster 2007), CLIL in Europe
appears to stress a dual focus on both language and content teaching (Coyle, Hood, and
Marsh 2010). This feature is essential in two perspectives. First, in some EFL educational
contexts where CLIL is implemented (e.g. some European and Asian countries), students
have limited exposure and opportunities to use the target language in their daily life.
Therefore, some explicit form-focused instruction will benefit students’ L2 development
(Llinares, Morton, and Whittaker 2012). Second, it has been found that the academic
language required to access academic content is different from the language that students
learn in language lessons or that they use in daily communication (Schleppegrell 2004).
Language is used to construct meaning and realise different cognitive processes in
specific ways in academic subjects (Leung 1996), which may not be transparent for
native speakers, not to mention L2 learners. Hence, incorporating more systematic
language teaching or at least raising students’ awareness of language use can facilitate L2
learning, on top of incidental learning via massive exposure to L2 input (Rose and
Martin 2012).
Despite the popularity of CLIL, it encounters several challenges, namely students’ L2
proficiency level on the onset of CLIL, teacher education, and curriculum and assessment
design (Mehisto, Marsh, and Frigols 2008). The first two aspects are directly relevant to
this paper and hence will be discussed in greater detail.
In his seminal work discussing factors affecting the effectiveness of bilingual
programmes, Cummins (1979) proposed that students have to attain a certain threshold
level of L1 and L2 proficiency when being admitted to those programmes. Otherwise,
272 Y.Y. Lo

they may suffer in their academic and cognitive achievements. As aforementioned, CLIL
is usually implemented in EFL contexts and it may be doubtful whether students
enrolling in the programme have reached the threshold level of English proficiency to
enjoy the benefits of CLIL. For instance, before 1997, over 90% of secondary schools in
Hong Kong claimed to be EMI schools where CLIL was practised (Tung, Lam, and
Tsang 1997), but it was found that only the top 40% of students would truly benefit from
and succeed in those schools (Education Department 1992). The issue may become more
severe considering the fact that most CLIL programmes start in secondary education,
where the academic knowledge has become increasingly abstract and cognitively more
difficult for L2 students. It is then more challenging for students to acquire the required
threshold level to access the academic content (Johnson and Swain 1994). As Coyle,
Hood, and Marsh (2010) put it, there may be a gap between students’ L2 and cognitive
levels.
The second major challenge to CLIL is teacher education. The preference for CLIL
creates a high demand for qualified teachers. Most CLIL teachers are non-native
speakers of the target language (Lasagabaster and Sierra 2010), and this has raised
some concerns about teacher education in CLIL. CLIL teachers not only need to be
knowledgeable in their specialist academic knowledge but they also need to be proficient
in the target language, as well as be able to scaffold students to learn the academic
language associated with the subjects ( Llinares, Morton, and Whittaker 2012). The job
of CLIL teachers is therefore very challenging and teacher preparation or in-service
training is the key to the effectiveness and sustainability of CLIL (Coyle, Hood, and
Marsh 2010).
One pedagogical implication of both the students’ limited L2 proficiency and teacher
education in CLIL is teachers’ language use in CLIL lessons. Facing students’ limited L2
standard and the abstractness of concepts in academic subjects, many CLIL subject
teachers have been observed to use L1 in order to avoid sacrificing students’ academic
achievements (e.g. García and Vazquez 2012; Wannagat 2007). This is regarded as a
‘realistic’ approach and ‘pragmatic’ solution to the learning problems caused by students’
limited L2 proficiency (Lin 2006). However, whether or not teachers should use L1 in
CLIL is still under debate.

Debate over L1 and L2 in CLIL

As previously noted, CLIL involves teaching of both L2 and content knowledge.
Therefore, the arguments over the use of L1 in CLIL classrooms represent a juxtaposition
of these two components. From the perspective of L2 learning, one major aim of CLIL is
to promote more effective L2 learning via greater exposure to L2 input. Such a principle
is in line with Krashen’s (1982) Input Hypothesis, which stresses the importance of plenty
of comprehensible language input for L2 acquisition. This hypothesis implicitly excludes
the use of L1 in classrooms. Also, under the influence of some second language teaching
methods (e.g. direct methods, audiolingualism), it occurs to some that different languages
should be kept separate so as to avoid negative transfer (Corder 1983).
Although the Input Hypothesis emphasises the importance of L2 input, it has been
argued that more L2 input does not necessarily imply more comprehensible L2 input, as
learners are often unable to process some L2 input effectively (Slimani 1992). This
problem is particularly acute in CLIL as L2 academic language and abstract content
knowledge impose considerable difficulties on L2 learners. Hence, it has been observed
 International Journal of Bilingual Education and Bilingualism 273

that teachers code-switch to L1 to provide translation equivalents for L2 academic terms,

or to provide examples from students’ daily life in order to explain difficult concepts (Lin
2006). In addition, it is also argued that using L1 can serve social and affective functions,
especially in CLIL classrooms where both teachers and students share the same L1
(Nikula 2007). It can also facilitate teacher–student interaction. When teachers receive no
responses to questions in L2, they are more likely to elicit students’ responses if they
code-switch to L1 (Then and Ting 2011). Moreover, contrary to the argument that
different languages should be kept apart, recent studies show that different languages
actually share some common underlying skills and processing capacity (Cummins 2000),
and the use of L1 can actually facilitate, instead of hindering, L2 learning (Jiang 2004).
For instance, in bilingual classrooms, teachers may highlight the difference(s) between L1
and L2, thereby fostering students’ metalinguistic awareness and L2 literacy development
(García and Vazquez 2012). L1 is also found to be an important tool to mediate students’
mental processes (e.g. tackling cognitively demanding problems, complex tasks; Swain
and Lapkin 2000). Therefore, rigidly maintaining the exclusive use of L2 is likely to
deprive students of a tool for mental processing and the pedagogy thus results would be
less cognitively demanding (McMillan and Turnbull 2009).
Given the potentials of using L1 in bilingual or CLIL classrooms, several researchers
have recently put forward ‘translanguaging’ pedagogies, which encourage the free use of
students’ multiple language repertoires to facilitate the learning processes and outcomes
in bilingual education (García, Flores, and Woodley 2012). However, some researchers
maintain that the use of L1 must be ‘limited’, ‘judicious’ and ‘theoretically principled’
(McMillan and Turnbull 2009). Yet, how should the word ‘judicious’ be defined? Macaro
(2005) suggested a threshold of 10–15% use of L1 in L2 learning classrooms, beyond
which the purpose of using L1 (to facilitate interaction and/or L2 learning) changes and
L2 learning may be adversely affected. However, the actual use of L1 varies considerably
in different bilingual classrooms. For instance, studies on Canadian immersion or
European CLIL classrooms revealed that teachers used none or very limited L1, and such
amount would decrease when students’ L2 develops (McMillan and Turnbull 2009;
Nikula 2007). On the other hand, studies on EMI classrooms in some post-colonial
countries or cities found teachers used as much as 85% of L1 in lessons with English
being the ‘official’ MOI (Probyn 2006; Wannagat 2007). Such a discrepancy can be
explained by the teachers, who may mediate between the official policy and classroom
practices (Li Wei and Martin 2009), based on their own L2 learning experiences and
students’ levels of abilities (McMillan and Turnbull 2009). However, teacher education or
training tends to play a less important role in teachers’ decision on language use, partly
because of a lack of such training and partly due to teachers’ own beliefs and perceptions
(McMillan and Turnbull 2009).
As the debate over ‘judicious’ use of L1 goes on, there are still concerns that once
teachers open the ‘floodgate’ of uncontrolled and excessive use of L1, it would be
difficult for them to keep the gate (Turnbull 2001). This may be a particular concern in
CLIL, considering the fact that content subject teachers usually prioritise delivering
subject knowledge and do not possess the same language awareness or perceived role of
language teaching as L2 teachers (Lo 2014). Will it be more difficult for CLIL teachers to
keep the ‘gate’? This study seeks to address this important question by investigating the
language use and teacher–student interaction patterns in several CLIL classrooms,
thereby inferring whether or not the teachers used L1 and L2 in a judicious way in
response to students’ needs and the dual goals of CLIL.
274 Y.Y. Lo

The overarching research question of this study is hence ‘To what extent do CLIL
teachers use languages (L1 and L2) in lessons?’, with the following specific questions:
(1) What is the pattern of language use in CLIL lessons?
(2) For what purposes do teachers use L1 and L2 in CLIL lessons?
(3) Do the amount and purposes of using different languages in CLIL lessons vary
according to students’ abilities and pedagogical needs?

Method
Research design
The data reported in this paper come from a larger-scale study investigating the classroom
interaction patterns and teachers’ language use in CLIL classrooms in Hong Kong (Lo
and Macaro 2012). This is a mixed-method observation study, employing a cross-
sectional design where the classroom interaction patterns of Grade 10 CLIL lessons were
observed and analysed.

Participants
The school
The participating schools in this study included five secondary schools, located in three
districts with similar demographics. These sampled schools were selected purposively
according to the MOI adopted, geographical location and students’ academic perform-
ance. This was to ensure that they were representative of Hong Kong secondary schools
and were also closely matched.
Under the government’s policy, Schools A, B and C were MOI-switching schools.
That is, they used Chinese as the MOI for Grades 7 to 9, but when students proceeded
to Grade 10, nearly all content subjects were taught in English, and students took public
examinations in English. On the other hand, Schools D and E were EMI schools
throughout – they adopted English as the MOI for nearly all subjects (except Chinese
history and literature) in all grades. The decision on the MOI was linked to the ability of
students admitted to the schools, and it is typical that students with good academic
results choose to enrol in EMI secondary schools. According to the information on the
school websites and the researcher’s conversation with the school principals and
teachers, among the five participating schools, the academic ability of students in the
two EMI Schools D and E was the highest (all Band 13 students). They were closely
matched by School A (with nearly all Band 1 students), then followed by School C
(with majority Band 1 and some Band 2 students) and School B (with majority Band 2
students). As a result, even though the ‘official’ MOI in Grade 10 for all the
participating schools was English, the teachers and students had different experiences
of CLIL and students were of varied academic abilities (refer to Table 1 for a summary
of the participating schools).

The teachers
Twelve teachers (seven females and five males) teaching content subjects (e.g. biology,
geography, history), identified as T1–T12, took part in this study. Their teaching
experience ranged from 3 to >20 years. There is no standardised English assessment for
content subject teachers in Hong Kong. The researcher did not ask the participating
 International Journal of Bilingual Education and Bilingualism 275

Table 1. Summary of the participating schools in the study.

School A School B School C School D School E

T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12

Characteristics
District located District A District B District A District C District A
MOI adopted Chinese in Grades 7–9; English in Grade English throughout Grades 7–13
10–13a
Student ability 100% Majority About 90% 100% 100% Band 1
Band 1b Band 2 + Band 1 + Band 1
some Band 1 some Band 2
and 3
a
Primary school leavers in Hong Kong are categorised into three bands, with Band 1 being the highest band.
b
The education system in Hong Kong has changed to 6-year secondary education (Grades 7–12) since 2012.

teachers about their English proficiency, and this constitutes a potential limitation.
However, from the lessons recorded and the illustrative samples provided, it is evident
that all teachers in the sample were adequately proficient in English to teach their subject
in Grade 10. The teachers were ensured that the lesson observations would only serve
research purpose so that they would conduct the lessons as usual, instead of altering their
teaching (especially their language use) for the researcher.

The students
Twelve classes of Grade 10 students (aged around 15–16 years) were involved in the
study, with approximately 480 students. The class size ranged between 38 and 42
students. All students used Chinese (or Cantonese to be more specific) as their mother
tongue. Students’ pass rate in English language in the public examination which all
students took at the end of Grade 11, was 99% in School A, 79% in School B, 97.4% in
School C, and 100% in Schools D and E. Such observable discrepancies in students’
standard of English across schools may help explain the differences, if any, in teachers’
language use. In addition, given the similar profile of Schools D and E, they are grouped
together (identified as EMI schools) when presenting and discussing the data in the
following sections.

Data collection
A total of thirty 35- to 40-minute lessons were observed in 20 visits. All lessons were
video- and/or audio-recorded. The researcher was always present during lesson
observations and additionally took field notes. Semi-structured interviews were also
conducted with the teachers and students. These interviews served to ensure that the
teachers and students behaved as usual in the lessons observed on one hand, and to
explore teachers’ and students’ opinions about the MOI policies and language use in
lessons. Most of the teachers and students regarded the observed lessons as normal,
thereby confirming the validity of the data collected. As teachers’ and students’ opinions
were out of the scope of this article, they are not reported in this paper.
276 Y.Y. Lo

Data analysis
Analysis of teachers’ language use
The lessons observed were analysed with several measures to present a picture of the
patterns of language use in the various schools.

Proportion of L1 and L2 use in lessons. Timed analysis was adopted to estimate the
proportion of teacher/student talk, as well as the proportion of L1 and L2 use. It was done
by sampling the lesson every 5 seconds and then coding it according to the speaker and
language being used (Macaro 2001).

Patterns of language use. If the teachers in the sampled schools were found to use a
noticeable proportion of L1, the lesson transcripts would be analysed in detail to examine
if there were certain patterns of using the two languages in the lessons. Lesson excerpts
would be presented to illustrate these patterns.

Functions of language use. Based on a taxonomy of functions (Table 3) derived from

previous studies (e.g. Canagarajah 1995; Lin 2006), the functions of the instances of L1
use found in the observed lessons were coded and the proportions of those different
functions in a lesson were calculated. The teachers in some schools were found to use L1
as the main MOI in the lessons, or code-switched between L1 and L2 frequently (see the
first two sections in Results). The purposes of using different languages in those schools
were then analysed separately.

Analysis of teacher–student interaction patterns

In addition to analysing teachers’ language use in lessons, the study also examined the
classroom interaction patterns in those lessons, to see if there was any potential
relationship between teachers’ language use, classroom interaction and the school
contexts. The classroom interaction patterns of the observed lessons have been reported
in greater detail elsewhere (Lo and Macaro 2012). For the purpose of this study, the
classroom interaction patterns were operationalised in two ways: first, proportion of
teacher and student talk, which was measured with Timed analysis as aforementioned;
second, mean lengths of teacher and student turns.

Results
Proportion of L1 and L2 use in lessons
The estimated proportions4 of L1 and L2 use are shown in Table 2.
Table 2 shows that teachers used different amounts of English and Chinese in
different schools, despite English being the ‘official’ MOI in Grade 10. Teachers in
Table 2. Proportion of L1 and L2 used in the observed lessons.

L1 use L2 use
School Mean % (SD) Mean % (SD)

A 20.7 (23.3) 79.3 (23.3)

B 80.2 (7.4) 19.8 (7.4)
C 61.9 (14.2) 38.1 (14.3)
D and E 96.8 (4.5) 3.26 (4.5)
 International Journal of Bilingual Education and Bilingualism 277

Schools D and E used English almost exclusively in their lessons (over 96% of
interaction time). Teachers in School A also attempted to stick to English, which occupied
around 80% of the interaction time in lessons. On the other hand, teachers in School B
used mainly Chinese, and so the proportion of English used was less than 20%. Teachers
in School C used about 60% Chinese and 40% English, roughly half L1 and half L2. If
one considers that students in Schools D and E were more able, followed by those in
School A, C and B, it seems that the proportion of L1 use in the observed lesson
negatively correlated with the students’ ability (i.e. teachers tended to use more L1 when
the students were weaker).

Patterns of L1 and L2 use

Since it was found that teachers in Schools A, B and C used at least 20% of L1 in their
lessons, it would be worth exploring the patterns of language use to see if there are any
differences.
In School A, teachers used English for most of the time, but on a few occasions, they
would switch to Chinese. This can be illustrated by lesson Excerpt 1, where the teacher
(T1) inserted one Chinese sentence (Line 8) during her explanation. (In all excerpts in this
paper, English is represented in capitals and Chinese in lower case; refer to the other
transcription conventions in the Appendix.)

Excerpt 1: (School A, biology)

This lesson was about food substances. T1 was trying to explain the results of emulsion test,
a test for fat/oil.
1 T1: … SO THE WATER MAKE THE FAT BECOME OIL DROPLET, OKAY, WHICH IS
VERY VERY TINY, SO IT BECOMES MILKY. BUT AFTER I ADD SOME WATER
INTO IT, NOW, ON TOP OF THE TUBE, OKAY, ON THE TOP OF THE SOLUTION,
WE CAN FIND A BIG DROPLET OF FAT. [pause; T demonstrating] NOT VERY CLEAR.
5 CAN YOU SEE THE SMALL CIRCLE IN THE MIDDLE?OKAY, THAT IS OIL,
OKAY, FLOATING ON THE TOP OF THE SOLUTION. OKAY, BECAUSE THE
WATER WILL DISPLACE THE FATS, OKAY, FROM THE ALCOHOL AND THE
FATTY MIXTURE. That means it separates the oil from the mixture. OKAY? BECAUSE
FATS CANNOT DISSOLVE IN WATER. IT ONLY DISSOLVES IN ETHANOL.
10 BUT WATER AND ETHANOL CAN MIX TOGETHER. OKAY?

In School B, teachers mainly used Chinese, but at the same time switched to English
quite often. This is because though they explained the content knowledge in Chinese,
they still needed to use the English technical terms of the subjects, such as ‘respiration’,
‘carbon dioxide’ and ‘energy’ in Excerpt 2.

Excerpt 2: (School B, biology)

This lesson was about respiration.

T3: Right, that means from this, from these two experiments, we know that mouse will
CARRY OUT RESPIRATION, then will PRODUCE CARBON DIOXIDE out. Is that right?
Then those (…) SEED, will also PRODUCE CARBON DIOXIDE. That means we know it
278 Y.Y. Lo

is one of the PRODUCT of RESPIRATION. Then, what are the other PRODUCT? You said
ENERGY just now. Then we have to talk about it in detail …

In School C, the teacher also used quite a lot of Chinese, but when he used English, it
would be sustained for a certain period of time, instead of the isolated patches
(representing only words or short phrases) as evident in School B. This can be
demonstrated in Excerpt 3, where T5 switched between Chinese and English sentences.

Excerpt 3: (School C, Biology)

The lesson was about food substances. The teacher was talking about carbohydrates.

T5: … OKAY, THE LAST ONE, POLYSACCHARIDES. POLYSACCHARIDES ARE IN

THE FORM OF A LONG CHAIN, LONG CHAIN OF MONOSACCHARIDES. MAYBE
SOME TYPES, IT IS THE BRANCH. It may have some branches. Then some branches
may link together. FOR EXAMPLE, STARCH IS AN EXAMPLE. IT IS A BRANCH
POLYSACCHARIDES, COMPOSED OF TWO, TWO DIFFERENT KINDS. There are
two kinds, two kinds of BRANCH. But concerning (…), there is no need to know.
CELLULOSE IS AN EXAMPLE, AND GLYCOGEN IS AN EXAMPLE. (pause)
OKAY, IN GENERAL, SUGAR IS WATER SOLUBLE, BUT EXCEPT FOR
POLYSACCHARIDES.

Functions of using different languages

This section presents the purposes of using English and Chinese in the participating
schools. As School A and the two EMI Schools D and E used English as the major
language of instruction in the lessons, they would be compared and discussed first. On
the other hand, Schools B and C used mainly Chinese and code-switched between the
two languages very frequently, and so they will be presented separately.

Purposes of using L1 in School A vs. Schools D and E

In both School A and Schools D and E, the teachers used English for most of the
interaction time in the lessons, but switched to Chinese occasionally. Based on the
taxonomy derived from previous studies, the proportions of different functions of code-
switches in a particular lesson were calculated from the total number of code-switches in
that lesson. This allows a comparison of the proportions of different functions of using L1
between School A and Schools D and E, which are presented in Table 3.
Table 3 reveals a similar pattern of the functions of code-switching in School A and
Schools D and E. Approximately, 60–70% of teachers’ code-switching was content-
related (i.e. to help deliver the content of the subjects to students), and around 30% of the
code-switches were used for classroom management to facilitate teaching and learning.
Very few code-switches served social or affective functions. There was only one instance
of T11 in School E using L1 to joke with the students.
One noticeable difference in the functions of using L1 lies in the pedagogical function
‘Explanation of difficult concepts’. About a quarter of the instances of code-switching in
School A served such a purpose, but there was none in Schools D and E. It was observed
that the teachers in School A switched to Chinese to explain difficult or abstract concepts
in lessons, as illustrated in Excerpt 4 (and also Excerpt 1 above).
 International Journal of Bilingual Education and Bilingualism 279

Table 3. Functions of teachers’ use of L1 in the observed lessons in School A and Schools D and E.

Major School A Schools D and E

categories Sub-categories Mean % Mean %

Social or 1. Referring to shared cultural norm or social 0 0

affective value
functions 2. Building up warmer and friendlier atmosphere 0 4.1
or building up rapport to students
(0) (4.1)
Pedagogical 3. Managing discipline 3.9 0
functions: 4. Comments on students’ behaviours (i.e. praise 0 18.5
Classroom and reprimand)
management 5. Giving instructions or commands 2.8 0.7
6. Encouraging class participation 2.8 0.4
7. Arousing students’ attention or focus 21.6 12.4
(30.8) (32.0)
Pedagogical 8. Explanation of difficult concepts 25.5 0
functions: 9. Parallel translation (i.e. providing the L1 43.6 58.8
Content equivalents for L2 vocabulary)
transmission 10. Providing annotations or examples in 0 5.2
students’ daily life to explain an unfamiliar topic
or concept to students
(69.1) (64.0)
Note: The percentages in brackets represent the subtotal percentage of a particular category.

Excerpt 4: (School A, geography)

This lesson was about fluvial processes.

T2: … OKAY, WHEN YOU LOOK AT EROSION AND TRANSPORTATION, YOU CAN
PAIR UP THEM. EROSION, WHEN ENERGY INCREASE, EROSION, THE FORCE OF
EROSION WILL INCREASE. MOREOVER, THE FORCE OF TRANSPORTATION
ALSO INCREASE. We have mentioned before, that energy flows from upper course to
lower course. If its energy is great, the river flow will be faster. When it is faster,
transportation, the transportation ability of river will be higher, higher. OKAY, DEPOS-
ITION, REMEMBER DEPOSITION? Deposition, deposition. OKAY, WHEN THE
ENERGY, RIVER ENERGY INCREASE, WHAT WILL HAPPEN TO THE DEPOSITION?

In Excerpt 4, the teacher first explained the positive relationship between river energy and
transportation. Then she switched to Chinese for a few sentences to explain that again. It
was obvious that using L1 here served the purpose of explaining difficult concepts, which
was often found in School A.
On the other hand, the teachers in Schools D and E seldom switched to Chinese for
explaining concepts. Instead, they tended to elaborate more in L2 to explain, or to interact
more with students so that students could grasp a better understanding of the concepts.
Excerpt 5 helps to illustrate such an attempt in Schools D and E.

Excerpt 5: (School D, history)

This lesson was about post-war recovery in Europe. The teacher was comparing the United
Nations and the League of Nations.
280 Y.Y. Lo

1 T8: WHEN WE TALK ABOUT MEMBERSHIP, BESIDES MEMBERS, YOU CAN

ALSO LOOK AT ANOTHER THING. BESIDES YOU TRIED TO FIND OUT WHO THE
MEMBERS ARE, YOU SHOULD ALSO THINK ABOUT THE ____? WHAT? (pause)
COMPARE, OKAY, LET’S COMPARE THE CLASS ASSOCIATION OF FOUR B, THE
5 CLASS ASSOCIATION OF FOUR B. HOW MANY MEMBERS DO YOU HAVE?
(pause) FOUR B, HOW MANY MEMBERS IN FOUR B? HOW MANY STUDENTS?
Ss: FORTY ONE.
T8: FORTY ONE, SO FORTY ONE MEMBERS. HOW ABOUT THE SA (Student
Association) (pause) OF (THE SCHOOL NAME)? YOU KNOW HOW MANY
10 STUDENTS ARE IN THE SCHOOL? ABOUT ____
Ss: ELEVEN …
T8: ELEVEN?
Ss: SA (Student Association)?
T8: SA, YES. SO HOW MANY STUDENTS ARE THERE IN (THE SCHOOL
NAME)?
15 (pause) AGAIN. YOU ARE WEAK IN MATH. DO YOU KNOW HOW TO DO IT?
FORM ONE, WE HAVE FIVE CLASSES. OKAY? SO EACH CLASS HAS ABOUT
____
S1: FORTY STUDENTS.
T8: FORTY STUDENTS, OKAY. DO THE CALCULATION QUICKLY.
S2: TWO HUNDRED.
20 T8: TWO HUNDRED? WHOLE SCHOOL!
S3: ONE THOUSAND AND SOMETHING.
T8: ONE THOUSAND AND SOMETHING. SOMETHING. ONE THOUSAND AND
ONE HUNDRED OR ONE THOUSAND.
Ss: (…)
25 T8: OKAY, OVER ONE THOUSAND. OKAY. SO WHAT CAN YOU SEE, WHEN
WE COME BACK TO THE UNITED NATIONS AND THE LEAGUE OF NATIONS?
(pause) WHICH ONE HAS MORE MEMBERS?
Ss: THE UNITED NATIONS.
T8: SO THE UNITED NATIONS HAD A ____? WIDER ____? (pause) MORE
30 MEMBERS, MORE ____? (pause) MORE MEMBERS, MORE ___?
S4: STORIES
T8: STORIES? I TOLD YOU LAST TIME. [pause; T wrote ‘RE’ on the blackboard]
Ss: RE … (pronounced as /ri/)
T8: RE ___? RE WHAT? NOT RE (/ri/).
S5: RESOLVED.
35 T8: RESOLVED? NO. MORE MEMBERS! SO FOR EXAMPLE, LET’S HAVE A
PICNIC TOMORROW, AND I JUST ASK THE CHAIRMAN AND THE VICE-
CHAIRMAN. OKAY? IT IS JUST LIKE THE CLASS ASSOCIATION. BUT IF I
ASKED THE WHOLE CLASS, MORE STUDENTS CAN MAKE THE DECISION.
SO WHAT CAN YOU SEE? (pause) WIDER ____? (pause) I CAN’T HEAR.
LOUDER. MORE ____?
40 Ss: REPRESENTATIVE … REPRE …
T8: THE DECISION IS NOT MADE BY ONE OR TWO PERSON!
Ss: REPRE … REPRESENTATIVE … REPRESENTATIVE …
T8: YES, RE ___
Ss: PRE …
45 T8: PRE
Ss: SEN …
T8: SEN
Ss: TA … TATIVE …
T8: REPRESENTA
50 Ss: TIVE … [pause; T wrote on the blackboard]
T8: OR YOU CAN SAY THE UNITED NATIONS HAS A WIDER REPRESENTA-
TION. OKAY? WIDER REPRESENTATION.
 International Journal of Bilingual Education and Bilingualism 281

In Excerpt 5, T8 was asking the students to compare the two peacekeeping bodies. At first,
the students did not get the idea that when talking about membership, one could focus on
both the actual number of members and the representativeness of the members. The teacher
tried to help by asking the students to consider their own ‘student association’ (Lines 5–
25). However, the students were still unable to grasp the point. Instead of resorting to
Chinese, the teacher interacted more with the students and provided a further hint (by
writing two letters on the blackboard, Lines 32–34) and an example (organising a picnic,
Lines 35–38). Finally, the students seemed to get the idea and the word ‘representative’,
and T8 then went even further by providing the nominalised form ‘representation’.
It was also observed that about half of the instances of using L1 in both School A and
Schools D and E served to provide parallel translation (i.e. providing the L1 equivalents
of English vocabulary so that students could grasp the meaning immediately). This
function was in fact commonly found in previous studies (e.g. Lin 2006), where the
researchers justified it by suggesting that it was more time-efficient and effective for
explaining the meanings of difficult vocabulary. Excerpt 6 is one of the many examples
found in the lessons observed.

Excerpt 6: (School D, History)

T8: … ACTUALLY WHAT DID THE USSR THINK? WHAT DID THE USSR THINK?
S1: SHE THINK IT WAS A CONSPIRACY.
T8: YES, RIGHT. SHE THOUGHT THAT IT WAS A CONSPIRACY. [pause; T writing on
the blackboard] DO YOU KNOW THIS WORD? DO YOU KNOW THIS WORD?
Conspiracy.

Purposes of code-switching in School B

Probably due to the lower English proficiency of the students, teachers in School B
maintained L1 as the MOI for most of the time in the lessons. Therefore, most code-
switching was from L1 to L2. As presented above, the purpose of code-switching to L2
was largely to teach English academic language (e.g. some subject-specific technical
terms or phrases). Apart from that major purpose, it was also observed that teachers code-
switched to L2 quite often to draw students’ attention to different aspects of L2 (e.g.
morphology, vocabulary, parts of speech and verb forms). Excerpt 7 is an example of T3
analysing morphology to help students understand or remember the subject-specific word
‘cellular’ more easily.

Excerpt 7: (School B, biology)

T3: … This one, so why do we call it CELLULAR RESPIRATION? It’s because of this. In
fact here, when you see this word, it looks unfamiliar, but actually we know it. Can you see,
we separate this part, this PREFIX CELL, it is ____?
Ss: Cells
T3: Cells. The meaning of CELLULAR here actually means connected to the cells.

In other words, teachers’ code-switching from L1 to L2 in School B sometimes

served the purpose of language teaching.
282 Y.Y. Lo

Purposes of code-switching in School C

The situations in School C were a mixture of those in Schools A and B. There, the
teachers taught in English for a while, usually at the beginning of the lessons, or when
talking about some factual information. Then they switched to Chinese with English
lexical items inserted when they had to explain something more difficult or abstract, or
try to elicit responses from the students. Therefore, there were code-switches from both
L2 to L1 and from L1 to L2, serving similar functions to those code-switches observed in
Schools A and B.
Similar to their counterparts in School B, teachers in School C also code-switched
from L1 to L2 to draw students’ attention to English. However, instead of explicitly
teaching different language aspects or skills, one interesting way of switching to L2
observed in School C was providing L2 translation as a form of ‘recast’, which is an
implicit form of corrective feedback (Lyster 2007) to students’ Chinese answers. This in
turn demonstrated the appropriate way of expressing the meaning in English to students.
Excerpt 8 shows how T6 of School C did that.

Excerpt 8: (School C, geography)

T6: When plates move, why will it shake? (pause) How can plates move?
S1: That way, that way, that way. [hands gesturing]
T6: That way, that way, that way. THEY MAY MOVE APART. (pause) OR ____? (pause) In
addition to moving apart, what are the other ways?
S2: Towards each other
T6: Towards each other, TOWARDS EACH OTHER. …

In Excerpt 8, T6 first accepted the students’ answer in Chinese. She then provided the
English phrases (e.g. towards each other) for the students. Compared with the explicit
language teaching shown in the previous excerpt, the language teaching in the form of
‘recast’ to students’ answers was more subtle.

Teacher–student interaction patterns

The proportions of teacher and student talk in the observed lessons are presented in
Table 4, whereas the mean length and number of teacher and student turns are shown in
Table 5.
All the observed lessons were indeed teacher-centred, as teacher talk occupied at least
90% of the interaction time and there were much more teachers’ turns than students’. The
lessons observed in School A seemed to be the most teacher-centred, as teacher talk
occupied over 95% of the interaction time, and the teachers on average held the floor for

Table 4. Proportion of teacher and student talk in the observed lessons.

Teacher talk Student talk

Schools Mean % (SD) Mean % (SD)

A 96.0 (1.7) 4.6 (1.8)

B 91.0 (4.1) 9.4 (4.3)
C 89.5 (4.2) 10.7 (4.0)
D and E 91.7 (4.4) 8.8 (4.2)
 International Journal of Bilingual Education and Bilingualism 283

Table 5. Total and mean length of teacher and student turns in the observed lessons.

Mean length of No. of teacher Mean length of No. of student

Schools teacher turns turns student turns turns

A 36.4 s (SD = 41.2) 226 2.6 s (SD = 2.0) 171

B 12.8 s (SD = 20.4) 873 1.8 s (SD = 1.1) 738
C 13.6 s (SD = 20.0) 622 1.9 s (SD = 1.6) 497
D and E 20.1 s (SD = 29.4) 652 2.3 s (SD = 2.2) 524

over 35 seconds before distributing the turn to students. On the other hand, lessons in
Schools B and C were more interactive as teachers there talked comparatively less and
their turns were considerably shorter than those in the other schools. Schools D and E
were somewhere in the middle between School A on one end and Schools B and C on the
other.
It is worth noticing that despite the fact that the statistics related to teacher talk/turns
show that lessons in School A were more teacher-centred, the mean length of student
turns there (2.6 seconds) was the longest among the schools. Nonetheless, it has been
found that those longer student turns consisted of quite a lot of fillers, pauses and false
starts, as students in School A mostly spoke in English. When student turns were
measured in syllables (so as to reflect the actual richness in the content), those in School
A were actually shorter than those observed in Schools D and E, where students also
mainly spoke in English (Lo and Macaro 2012).
The findings about teachers’ language use and classroom interaction patterns, together
with the information about student ability, are summarised in Figure 1. In School A, the
teachers tried to maintain English as MOI, and they sometimes code-switched to Chinese
to explain difficult concepts. The lessons there appeared to be very teacher-centred. In
School B, teachers mainly used Chinese, with English academic terms inserted. More
teacher–student interaction was observed. In School C, teachers switched between
English and Chinese, and similar to School B, more teacher–student interaction was
found. Finally, in Schools D and E, the teachers largely used English in the lessons,
sometimes complemented with Chinese translations of difficult English words. The
lessons there appeared to be more interactive than those in School A but less so than
those in Schools B and C.

Figure 1. Summary of findings.

284 Y.Y. Lo

Discussion
This study attempted to explore whether and how teachers made use of different linguistic
resources to help students overcome the potential barriers of learning L2 and academic
content simultaneously in CLIL in the Hong Kong context. It did not seek to evaluate the
schools, teachers or students, and the findings are by no means an indicator of the quality
or effectiveness of the schools.
One major conclusion from the findings is that teachers’ language use in lessons was
framed by several factors. The first one is the MOI policy adopted by the school. The
influence of this factor is particularly prevalent in Schools D and E, the EMI schools
throughout, and School A, a school switching to English as the MOI in Grade 10. In
those schools, teachers appeared to be aware of the need to maintain the ‘official’ MOI,
even though that may result in limited interaction between teachers and students (as in
School A). Such a pressure to conform to the government or school policy, as well as the
expectations of stakeholders (e.g. parents) is similar to other contexts like Canada
(McMillan and Turnbull 2009).
However, the official policy was sometimes mediated by other factors, especially
students’ English proficiency, when teachers chose the actual language of instruction in
lessons (Li Wei and Martin 2009; Probyn 2006). In Schools A, D and E, teachers largely
adhered to the official MOI policy probably because the students there were more able
and more proficient in English. They at least could understand the academic content
delivered in English, though some may not be able to respond or participate in interaction
actively. In contrast, students in Schools B and C, particularly the former, possessed
lower English proficiency, and hence teachers there spoke in Chinese rather often.
Otherwise, students may not be able to understand the academic content, thereby
sacrificing their academic achievement in CLIL.
In addition to the overall language pattern that teachers followed in their lessons, this
study also observed some different functions of L1 and L2 in the various school contexts.
While teachers in Schools A, D and E code-switched to Chinese to manage the class, to
translate English words and to explain abstract concepts, their counterparts in School B
code-switched to English to do some form-focused teaching, and teachers in School C
used English to provide ‘recasts’ to students’ Chinese utterances, thereby teaching the
target language at the same time. In this sense, it may be reasonable to argue that CLIL
teachers, despite the school contexts they were working in, were sensitive to adopt
different languages to try to achieve both content and language learning, given the school
policy and students’ learning needs.
Though it seems that teachers’ language use took into consideration their contexts and
students, some teachers in this study did appear to use quite a large amount of L1. For
example, the teachers in School B used around 80% of Chinese in the lessons, which far
exceeded the 10–15% threshold proposed by Macaro (2005). Such practices resemble
what García, Flores, and Woodley (2012) called ‘translanguaging’ in bilingual education.
However, there may be concerns whether students would be able to express their
understanding of content knowledge through English academic language, which has been
shown to be significantly different from everyday communicative language (Schleppe-
grell 2004).
Thus, for schools which implement CLIL but where students are not sufficiently
equipped with L2 proficiency, what are some implications of this study?
First, the school may need to provide extra support to enhance students’ L2
proficiency, especially related to academic language. This can be done by providing
 International Journal of Bilingual Education and Bilingualism 285

some small-class intensive bridging courses before or at the beginning stage of CLIL, if
school resources permit (Johnson and Swain 1994). If students can follow what teachers
say in L2 in lessons and participate in interaction, teachers will in turn be less inclined to
use L1, and hence strike a more proper balance between content and language learning.
Second, teachers can try encouraging or maintaining teacher–student interaction in L2
as much as possible, and perhaps treat L1 as the last resort. Facing the tight syllabus and
time constraint, content subject teachers were observed to switch to L1 quite frequently
for explanation. Sometimes, they also seemed to get frustrated by no student responses or
limited interaction. However, some strategies observed in Schools D and E in this study
appeared to facilitate L2 interaction. For instance, the teachers provided some time for
students’ individual, pair or group work before nominating students to give responses.
Such a ‘Think, Pair, Share’ strategy is very common in language lessons and can be
adopted in CLIL content subject lessons as well (Tavares’ paper in this volume also
demonstrates this). Also, teachers were found to engage students in prolonged
interactional exchanges (e.g. in Excerpt 5) to explore difficult concepts, instead of
switching to L1. This was achieved by teachers’ skills in asking and modifying questions,
such as providing clues, using Socratic questioning, making the topic more salient, etc.
(Tsui 1995).
Third, even though the teachers do code-switch to L1 for pedagogical purposes, it
might be better if they would reinforce the concepts in L2 again. As observed in School
C, the teachers sometimes discussed difficult concepts or elicited students’ responses in
L1. Yet, they would follow that up with L2 explanations or illustration. This not only
serves as reiteration or reinforcement of the concepts but also a kind of feedback to
facilitate students’ language learning.
Therefore, this study yields an important implication for teacher education in CLIL –
teachers should be aware of their role in enabling students to make the best use of their
linguistic repertoire so as to move strategically between L1, everyday L2 and academic
L2, instead of resorting primarily to L1 most of the time (see the Rainbow diagram
proposed by Lin 2012, 93; also in the Introduction paper of this volume).

Conclusion
The spread of CLIL to EFL contexts, where students enrolling in CLIL may not be
equipped with sufficient English proficiency, creates a dilemma for teachers – using L1
can help students better understand the abstract content knowledge, but it reduces L2
input and may go against the rationale of CLIL to facilitate L2 learning. Hence, whether
teachers should use L1 in CLIL lessons is still an unresolved issue. Though researchers
have suggested ‘judicious’ use of L1, it remains unclear how to define ‘judicious’. Should
it be defined as a certain threshold percentage out of the lesson time? Or should it also
take into account the diverse contexts, teachers’ and students’ needs? The 30 lessons
observed in different CLIL classrooms in Hong Kong in this study show that teachers
were rather sensitive in their language use, considering the MOI policy, students’ L2
proficiency and academic ability. Nevertheless, whether such use of L1 can help students
grasp the content knowledge and whether students can then express their understanding
through academic L2 remains unexplored. Future research may examine if there is any
trade-off between using L1, access to content knowledge and L2 learning in CLIL
lessons.
286 Y.Y. Lo

Notes
1. As discussed in the Introduction paper, CLIL in this paper (and throughout this special issue) is
used as an umbrella term to incorporate a variety of programmes which use L2 as the MOI for
content subjects. Hence, this paper includes references to the research on different variants of
CLIL implemented in different contexts (e.g. CLIL implemented in Europe and immersion
programmes in Canada). Although it is acknowledged that these variants may be different in a
number of ways, as Cenoz, Genesee, and Gorter (2014) argue, referring to relevant literature in
various educational setting can further enhance the effectiveness and practice of CLIL.
2. Code-switching is broadly defined as the use of more than one language by a single speaker in
the course of a conversation. From the second language acquisition perspective, the
phenomenon of code-switching often concerns the use of L1 when the teacher shares the
same mother tongue with the students in classrooms where the L2 is the default medium of
teaching. However, as the research later unfolded, some teachers did frequently alternate
between L1 and L2, and so the term ‘code-switching’ is also used in this study.
3. Band 1 is the highest band in the three-tier categorisation system of Primary 6 school-leavers in
Hong Kong.
4. The percentages shown in this paper, unless otherwise specified, represent the percentage of
time out of the interaction time (i.e. total lesson time minus non-interaction time).

References
Canagarajah, A. S. 1995. “Functions of Code Switching in the ESL Classroom: Socialising
Bilingualism in Jaffna.” Journal of Multilingual and Multicultural Development 16 (3): 173–
196. doi:10.1080/01434632.1995.9994599.
Cenoz, J., F. Genesee, and D. Gorter. 2014. “Critical Analysis of CLIL: Taking Stock and Looking
Forward.” Applied Linguistics 35: 243–262.
Corder, S. 1983. “A Role for the Mother Tongue.” In Language Transfer in Language Learning,
edited by S. Gass and L. Selinker, 85–97. Rowley, MA: Newbury House.
Coyle, D., P. Hood, and D. Marsh. 2010. CLIL: Content and Language Integrated Learning.
Cambridge: Cambridge University Press.
Cummins, J. 1979. “Linguistic Interdependence and the Educational Development of Bilingual
Children.” Review of Educational Research 49 (2): 222–251. doi:10.3102/003465430
49002222.
Cummins, J. 2000. Language, Power, and Pedagogy: Bilingual Children in the Crossfire.
Clevedon: Multilingual Matters.
Dalton-Puffer, C., T. Nikula, and U. Smit. 2010. “Language Use and Language Learning in CLIL:
Current Findings and Contentious Issues.” In Language Use and Language Learning in CLIL
Classrooms, edited by C. Dalton-Puffer, T. Nikula, and U. Smit, 279–291. Amsterdam: John
Benjamins.
Education Department. 1992. A Study on the Relation between Initial Language Proficiency at S1
Level and Subsequent HKCEE Performance for Medium-of-Instruction Grouping. Hong Kong:
Government Printer.
García, M. C. M., and V. P. Vazquez. 2012. “Investigating the Coexistence of the Mother Tongue
and the Foreign Language through Teacher Collaboration in CLIL Contexts: Perceptions and
Practice of the Teachers Involved in the Plurilingual Programme in Andalusia.” International
Journal of Bilingual Education and Bilingualism 15 (5): 573–592.
García, O., N. Flores, and H. Woodley. 2012. “Transgressing Monolingualism and Bilingual
Dualities: Translanguaging Pedagogies.” In Harnessing Linguistic Variation to Improve
Education, edited by A. Yiakoumetti, 45–75. Bern: Peter Lang.
Jiang, N. 2004. “Semantic Transfer and Its Implications for Vocabulary Teaching in a Second
Language.” Modern Language Journal 88 (3): 416–432. doi:10.1111/j.0026-7902.2004.00
238.x.
Johnson, R. K., and M. Swain. 1994. “From Core to Content: Bridging the L2 Proficiency Gap in
Late Immersion.” Language and Education 8 (4): 211–229. doi:10.1080/09500789409541392.
Krashen, S. 1982. Principles and Practice in Second Language Acquisition. New York:
Prentice Hall.
Lasagabaster, D., and J. M. Sierra. 2010. “Immersion and CLIL in English: More Differences than
Similarities.” ELT Journal 64: 367–375. doi:10.1093/elt/ccp082.
 International Journal of Bilingual Education and Bilingualism 287

Leung, C. 1996. “Context, Content and Language.” In Curriculum Related Assessment: Cummins
and Bilingual Children, edited by T. Cline and N. Frederickson, 26–40. Clevedon: Multilingual
Matters.
Li Wei, and P. Martin. 2009. “Conflicts and Tensions in Classroom Codeswitching: An
Introduction.” International Journal of Bilingual Education and Bilingualism 12 (2): 117–
122. doi:10.1080/13670050802153111.
Lin, A. M. Y. 2006. “Beyond Linguistic Purism in Language-in-education Policy and Practice:
Exploring Bilingual Pedagogies in a Hong Kong Science Classroom”. Language and
Education 20: 287–305. doi:10.2167/le643.0.
Lin, A. M. Y. 2012. “Multilingual and Multimodal Resources in Genre-based Pedagogical
Approaches to L2 English Content Classrooms.” In English – A Changing Medium for
Education, edited by C. Leung and B. V. Street, 79–103. Bristol: Multilingual Matters.
Llinares, A., T. Morton, and R. Whittaker. 2012. The Roles of Language in CLIL. Cambridge:
Cambridge University Press.
Lo, Y. Y. 2014. “Collaboration between L2 and Content Subject Teachers in CBI: Contrasting
Beliefs and Attitudes.” RELC Journal 45 (2): 181–196.
Lo, Y. Y., and E. S. C. Lo. 2014. “A Meta-analysis of the Effectiveness of English-medium
Education in Hong Kong.” Review of Educational Research 84 (1): 47–73. doi:10.3102/
0034654313499615
Lo, Y. Y., and E. Macaro. 2012. “The Medium of Instruction and Classroom Interaction: Evidence
from Hong Kong Secondary Schools.” International Journal of Bilingual Education and
Bilingualism 15 (1): 29–52. doi:10.1080/13670050.2011.588307.
Lyster, R. 2007. Learning and Teaching Languages through Content: A Counterbalanced
Approach. Amsterdam: John Benjamins.
Macaro, E. 2001. “Analysing Student Teachers’ Codeswitching in Foreign Language Classrooms:
Theories and Decision Making.” Modern Language Journal 85 (4): 531–548. doi:10.1111/
0026-7902.00124.
Macaro, E. 2005. “Codeswitching in the L2 Classroom: A Communication and Learning Strategy.”
In Non-native Language Teachers: Perceptions, Challenges, and Contributions to the
Profession, edited by E. Llurda, 63–84. Boston, MA: Springer.
McMillan, B., and M. Turnbull. 2009. “Teachers’ Use of the First Language in French Immersion:
Revisiting a Core Principle.” In First Language Use in Second and Foreign Language Learning,
edited by M. Turnbull and J. Dailey-O’Cain, 15–34. Bristol: Multilingual Matters.
Mehisto, P., D. Marsh, and M. J. Frigols. 2008. Uncovering CLIL: Content and Language
Integrated Learning in Bilingual and Multilingual Education. Oxford: Macmillan Education.
Nikula, T. 2007. “Speaking English in Finnish Content-based Classrooms.” World Englishes 26 (2):
206–223. doi:10.1111/j.1467-971X.2007.00502.x.
Probyn, M. J. 2006. “Language and Learning Science in South Africa.” Language and Education
20 (5): 391–414. doi:10.2167/le554.0.
Rose, D., and J. Martin. 2012. Learning to Write, Reading to Learn: Genre, Knowledge and
Pedagogy in the Sydney School. Sheffield: Equinox.
Schleppegrell, M. 2004. The Language of Schooling: A Functional Linguistics Perspective. New
York: Routledge.
Slimani, A. 1992. “Evaluation of Classroom Interaction.” In Evaluating Second Language
Education, edited by J. C. Alderson and A. Beretta, 197–221. Cambridge: Cambridge
University Press.
Swain, M. 1986. “Two Ingredients to the Successful Use of a Second Language as a Medium of
Instruction in Hong Kong.” Educational Research Journal 1 (1): 1–6.
Swain, M., and S. Lapkin. 2000. “Task-based Second Language Learning: The Uses of the First
Language.” Language Teaching Research 4 (3): 251–274. doi:10.1177/136216880000400304.
Then, D. C. O., and S. H. Ting. 2011. “Code-switching in English and Science Classrooms: More
than Translation.” International Journal of Multilingualism 8 (4): 299–323. doi:10.1080/
14790718.2011.577777.
Tsui, A. B. M. 1995. Introducing Classroom Interaction. London: Penguin.
Tung, P., R. Lam, and W. K. Tsang. 1997. “English as a Medium of Instruction in Post-1997 Hong
Kong: What Students, Teachers and Parents Think.” Journal of Pragmatics 28 (4): 441–459.
doi:10.1016/S0378-2166(97)00034-9.
288 Y.Y. Lo

Turnbull, M. 2001. “There is a Role for the L1 in Second and Foreign Language Teaching, but …”
Canadian Modern Language Review 57 (4): 531–540.
Turnbull, M., and J. Dailey-O’Cain, eds. 2009. First Language Use in Second and Foreign
Language Learning. Bristol: Multilingual Matters.
Wannagat, U. 2007. “Learning through L2 – Content and Language Integrated Learning (CLIL) and
English as Medium of Instruction (EMI).” International Journal of Bilingual Education and
Bilingualism 10 (5): 663–682.

Appendix. Transcription conventions

T1, T2, etc. = Teacher 1, Teacher 2, etc.
S1, S2, etc. = single student.
Ss = more than one student.
(…) = inaudible utterances.
(italics) = words added to make the utterances comprehensible.
[] = non-verbal actions or author’s comments.
___ (at the end of questions) = short pauses indicating blank filling questions.